This is Prog-09.gz in view mode; [Up]
From: kcd@babylon5.jumpgate.com (Kenneth C. Dyke) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 30 Aug 1997 23:43:03 GMT Message-ID: <5uab67$fgp$1@nntp2.ba.best.com> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <5ua3k8$nmt$1@owl.slip.net> In-Reply-To: <5ua3k8$nmt$1@owl.slip.net> On 08/30/97, Steve Dekorte wrote: > >This is all correct, but... >As long as the subclasses use accessor methods for all variable references, >there should be no need to recompile even subclasses. Actually, if a subclass has it's own instance variables, you will run the risk of accidentally overwriting your superclass's instance variables if some got added and you don't recompile. Likewise, the superclass would potentially be overwriting your instance variables as well. -Ken -- Kenneth Dyke, kcd@jumpgate.com (personal), kdyke@ea.com (work) Nuclear Strike and OPENSTEP Tools Engineer, Electronic Arts C++: The power, elegance and simplicity of a hand grenade.
From: kennel@nospam.lyapunov.ucsd.edu (Matt Kennel (Remove 'NOSPAM' to reply)) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: DPS and threading (was Re: MacOS 8 catching up to Windows 95?) Date: 31 Aug 1997 01:00:57 GMT Organization: The Avant-Garde of the Now, Ltd. Message-ID: <slrn60hghr.su2.kennel@lyapunov.ucsd.edu> References: <Pine.SOL.3.91.970827225446.3629A-100000-100000-100000-100000-100000-100000-100000-100000-100000@ux4.cso.uiuc.edu> <joe.ragosta-ya02408000R2808970813220001@news.dol.net> <Pine.SOL.3.91.970828161940.4553A-100000@ux5.cso.uiuc.edu> On Thu, 28 Aug 1997 16:26:49 -0500, Ryan Tokarek <tokarek@uiuc.edu> wrote: :> I believe Rhapsody will do the same thing (unless your program bypasses :> DPS). Your app will have its own thread and DPS will have another. So, at :> the very least, you get two threads per app. : :I don't think this is right (though I could easily be wrong). Actually, :I'm not all that clear on all the DPS/application/WindowServer business. : :Could somebody explain roughly how threading works with respect to all :this in NeXT/OpenStep? The window server is an entirely different process from all applications and draws asynchronously to the display screen. And inside the window server there may be multiple threads, but this is handled separately from application drawing. From all indications, there will be no performance problems with Rhapsody. Unless you really need to have 8-16 CPU's simultaneously doing calculations for one single on-screen render (e.g. like an SGI) there is no issue for multiprocessing under OpenStep. :Ryan Tokarek :<tokarek@uiuc.edu> -- * Matthew B. Kennel/Institute for Nonlinear Science, UCSD - * "People who send spam to Emperor Cartagia... vanish! _They say_ that * there's a room where he has their heads, lined up in a row on a desk... * _They say_ that late at night, he goes there, and talks to them... _they *- say_ he asks them, 'Now tell me again, how _do_ you make money fast?'"
From: sanguish@digifix.com Newsgroups: comp.sys.next.advocacy,comp.sys.next.announce,comp.sys.next.hardware,comp.sys.next.software,comp.sys.next.misc,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.sys.next.programmer Subject: NEXTSTEP/OpenStep Resources on the Net Supersedes: <10179872395226@digifix.com> Date: 31 Aug 1997 03:53:18 GMT Organization: Digital Fix Development Message-ID: <16621873000033@digifix.com> Topics include: Major OpenStep/NEXTSTEP World Wide Web Sites OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups Major OpenStep/NEXTSTEP FTP sites NeXTanswers Major OpenStep/NEXTSTEP World Wide Web Sites ============================================ The following sites are a sample of the OpenStep related WWW sites available. A comprehensive list is available on Stepwise. Stepwise OpenStep/Rhapsody Information Server http://www.stepwise.com Stepwise has been serving the OpenStep/NEXTSTEP community since March 1993. Some of the many resources on the site include: OpenStep Third Party Software guide, Developer Directory, Mailing List information, extensive listing of FTP and WWW sites related to OpenStep and NEXTSTEP, OpenStep related Frequently Asked Questions. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. This is the World Wide Web interace to the FTP site. Apple Enterprise Software Group (formerly NeXT Computer, Inc.) http://www.next.com Here is where you'll find the NeXTanswers archive, with information on OpenStep installation, drivers and software patches. Apple Computer's 'Prelude to Rhapsody' Self Support Site http://devworld.apple.com/dev/prelude.html This site has been constructed to help you help yourself to learn as much as possible about the foundation for Rhapsody, today's OPENSTEP. The site provides an informal collection of pointers, references, and starting points for developers who are using the Prelude to Rhapsody bundle, distributed at this year's Worldwide Developer Conference. OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups ==================================================== COMP.SYS.NEXT.ADVOCACY This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. COMP.SYS.NEXT.ANNOUNCE Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. COMP.SYS.NEXT.BUGS A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT-specific groups as well. COMP.SYS.NEXT.HARDWARE Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. COMP.SYS.NEXT.MARKETPLACE NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. COMP.SYS.NEXT.MISC For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! COMP.SYS.NEXT.PROGRAMMER Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. COMP.SYS.NEXT.SOFTWARE This is a place to talk about [third party] software products that run on NEXTSTEP systems. COMP.SYS.NEXT.SYSADMIN Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. ** RELATED NEWSGROUPS ** COMP.SOFT-SYS.NEXTSTEP Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. COMP.LANG.OBJECTIVE-C Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. COMP.OBJECT Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep next-advocacy next-announce next-bugs next-hardware next-marketplace next-misc next-programmer next-software next-sysadmin object lang-objective-c (For a full description, send mail to listserv@antigone.com). The subscription syntax is essentially the same as Majordomo's. To subscribe, send a message to *-request@lists.best.com saying: subscribe where * is the name of the list e.g. next-programmer-request@lists.best.com Major OpenStep/NEXTSTEP FTP sites ================================= ftp://ftp.next.peak.org The main site for North American submissions formerly ftp.cs.orst.edu ftp://ftp.peanuts.org: (Peanuts) Located in Germany. Comprehensive archive site. Very well maintained. ftp://ftp.nluug.nl/pub/comp/next NeGeN/NiNe (NEXTSTEP Gebruikers Nederland/NeXTSTEP in the Netherlands) ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! _________________________________________________________________ Written by: Eric P. Scott ( eps@toaster.SFSU.EDU ) and Scott Anguish ( sanguish@digifix.com ) Additions from: Greg Anderson ( Greg_Anderson@afs.com ) Michael Pizolato ( alf@epix.net ) Dan Grillo ( dan_grillo@next.com )
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: c.l.objc FAQ update (Re: Objective-C Fragile Base Class) Date: 31 Aug 1997 19:39:04 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7sovqi75z.fsf_-_@tom.ics.ele.tue.nl> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <EFqDnH.I15@basil.icce.rug.nl> In-reply-to: tom@basil.icce.rug.dev.null.nl's message of 30 Aug 1997 16:42:22 GMT In article <EFqDnH.I15@basil.icce.rug.nl> tom@basil.icce.rug.dev.null.nl (Tom Hageman -- remove .dev.null to reply) writes: Does it support categories? (the ObjC FAQ doesn't tell...) In response to this question (and with a Sunday of free time to spare) I have revised part of the comp.lang.Objective-C FAQ. Question 17 in ftp://ftp.ics.ele.tue.nl/pub/objc/answers addresses the differences between the various compilers and runtimes. As always, remarks and other feedback are appreciated. --Tiggr
From: ckane@next.com (Christopher Kane) Newsgroups: comp.sys.next.programmer Subject: Re: virtual retains ?? Date: 31 Aug 1997 18:04:13 GMT Organization: Apple Computer, Inc. Distribution: world Message-ID: <5ucbmt$hfk$1@news2.apple.com> References: <5u1p7f$18g$1@concorde.ctp.com> In article <5u1p7f$18g$1@concorde.ctp.com> tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) writes: > while running one of my apps from ObjectAlloc I received these "logs" > from my application. However..when running it under regular conditions > nothing shows up. > > The message is: > Aug 27 18:44:00 DocumentManagement[804] *** _NSNegativeRetain(): object > (0xc702cc) of class DBDocumentContainer has less than zero (-1) virtual > retains The "virtual retain count", as refered to by this message, is simply the -retainCount minus the number of pending autoreleases on the object (in the current thread only, in 4.2). The virtual retain count being less than 0 means that if all of those autoreleases turn into releases, by the pools being freed, one of the -release messages will eventually be sent to a freed object. NSDebug.h talks about this, I believe. A virtual retain count less than 0 indicates some programming error, but you may "get away with it". The object may be in some top-level pool which is never freed (and so the autorelease->release never happens). The object may be doing its own reference counting incorrectly. The object may be retained again before all of the pool's it is in are released. Or something else. Christopher Kane Application Frameworks Apple Computer, Inc.
From: ckane@next.com (Christopher Kane) Newsgroups: comp.sys.next.programmer Subject: Re: Storage Class gone?? Date: 31 Aug 1997 18:15:32 GMT Organization: Apple Computer, Inc. Distribution: world Message-ID: <5uccc4$vu4$1@news2.apple.com> References: <EXMckvKmrW$b@cc.usu.edu> In article <EXMckvKmrW$b@cc.usu.edu> writes: > I'm in the process of converting several apps from NextStep > to OpenStep. It would appear that the Storage Class has > disappeared from the environment. > > Is there a way to coax any FoundationKit classes into > performing in a similar fashion? I'd just like a class which > will manage a collection of structures, like the Storage > class used to let me. I use an NSMutableData object, and define a couple convenience access macros. Strictly speaking, Storage is still available in 4.x (it's still <objc/Storage.h>). I'd encourage you to convert off of it though, as it might disappear more completely, except for unarchiving purposes, in Rhapsody or beyond. Christopher Kane Application Frameworks Apple Computer, Inc.
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <81f7cd$10271a.3d6@NEWS> Control: cancel <81f7cd$10271a.3d6@NEWS> Date: 31 Aug 1997 21:26:59 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.81f7cd$10271a.3d6@NEWS> Sender: grants@ForYou.com Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: scholz@leo.org (Bernhard Scholz) Newsgroups: comp.soft-sys.nextstep,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,pgh.next-users,tw.bbs.comp.os.nextstep Subject: Re: Tar.gz : how can I expand this files ? Date: 31 Aug 1997 20:45:40 GMT Organization: [posted via] Leibniz-Rechenzentrum, Muenchen (Germany) Distribution: world Message-ID: <5ucl5k$5g4$1@xeres.hsh.stusta.mhn.de> References: <01bcb3d8$02388e60$050e03c3@worldnet.sct.fr> "Laurent" <solfra@worldnet.fr> wrote: >Hello next-web afficionados, > > >I don't know how I can expand the files with the extension b.tar.gz with >Openstep for Mach 4.1. > The ending .b. indicates it's a binary and is of no further use. tar indicates 'TAR' archives and gz indicates 'GnuZIP. You can use gunzip and tar -x together or use gnutar -zxf <file> Greetings, Bernhard. -- Bernhard Scholz http://www.leo.org/~scholz/ Peanuts FTP Admin http://peanuts.leo.org/ scholz@leo.org, (StuSta ONLY: boerny@xenia.hsh.stusta.mhn.de)
Newsgroups: comp.sys.next.programmer Subject: Comparing NS Development Time Message-ID: <lukas-3008972357410001@lanrower2-3.urz.unibas.ch> From: lukas@iso.iso.unibas.ch (Lukas Schuepbach) Date: Sat, 30 Aug 1997 23:57:41 +0200 I am thinking about starting a new software project. I have already made several projects with METROWERKS C++ and I would like to know what developement time I have to expect compared to the Mac OS when I am developing for NS. Thank's for any help Lukas
From: Nick@tzcc.com (Nick Dinielli (The Fatman)) Newsgroups: comp.sys.next.programmer Subject: http://www.tzcc.com/ all you need to know for custom applications and programs. Date: Mon, 01 Sep 1997 03:01:22 GMT Organization: NetTrip, Inc. Message-ID: <32200e5f.179100975@news.carroll.com> Mime-Version: 1.0 Content-Type: text/plain http://www.tzcc.com/ all you need to know for custom applications and programs.
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <32200e5f.179100975@news.carroll.com> Control: cancel <32200e5f.179100975@news.carroll.com> Date: 01 Sep 1997 03:15:31 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.32200e5f.179100975@news.carroll.com> Sender: Nick@tzcc.com (Nick Dinielli (The Fatman)) Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: Josep Giribet <calidoscopi@redestb.es> Newsgroups: comp.sys.next.programmer Subject: Re: Rhaposody DR1 in Apple Dev. Program Date: Sun, 31 Aug 1997 17:51:03 +0100 Organization: CALIDOSCOPI Message-ID: <3409A0F8.1CC6@redestb.es> References: <33F8B9AB.99F@anonym.at> <33f8fe1e.0@news.intnet.net> <33FC8B50.58BE@gibbsx.oit.unc.edu> <33FDCB77.66D5@emory.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Curtis Crowson wrote: > Is it still slated to work only on 85/8600 and 95/9600 machines?? > > I'd do it in a heartbeat if it'd work on my 7200. But it's cheaper for > me to buy an Intel clone ($750 P-133) and install NeXTStep ($300 edu) > than it is to buy an 8600. True, and I have a 9600 at work but it's busy all day rendering (we are actually doing work on it, we can't really afford to have another 9600 to develop out of our current projects), so I'll have to install it on our Intel backup machine :p.
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.next.programmer Subject: OpenStep programming for newbies Date: 1 Sep 1997 14:31:51 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5udgfn$t7o$1@attila.apana.org.au> I have a project coming up, and I'm trying to select tools for the job. I've been curious about OpenStep for years, and figure that this might be a good opportunity to actually do something with it. I like to think that I'm familiar with OO concepts (started with ObjectPascal, then C++ and Java), and I've learnt enough different languages to be able to pick up a new one without too much trouble. There are two questions I need to answer before doing this on OpenStep, though: * How long should it take me to get up to speed? Would it be reasonable to expect to be able to do this project (more details below) after a couple of weeks teaching myself OpenStep? * How suitable is OpenStep for the job? It's a point-of-sale and stock tracking system. The main issues for me are database services (ideally I want to put the database server on one beefy machine, and have smaller machines to run the client on); keyboard control (the customer, quite reasonably, wants to be able to do everything with a keyboard); and cost. Any help would be greatly appreciated. Matt -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: "Eric A. Dubiel" <eadubie@rs6000.cmp.ilstu.edu> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.mac.advocacy Subject: "The OpenStep Specification" Date: Mon, 01 Sep 1997 01:23:58 -0600 Organization: Illinois State University Message-ID: <340A6D61.6A88@rs6000.cmp.ilstu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit The following message was posted on the GNUstep mailing list not so long ago, and there was never a satisfactory answer, can anyone comment on this who has some authoritative knowledge. Perhaps "Apple Computer, NeXT Campus" employees?! ;-) Thanks! -Eric What is to become of "The OpenStep Specification", now that Apple is apparently modifying and expanding it in the process of building Rhapsody? Will all components of the Yellow Box be considered part of the "new and improved" OpenStep? Does Sun still have any say in its evolution? Has it been submitted to a standards body, as was once promised? If not, are there still plans to do so? --Eric --- Eric K. Ringger mailto:ringger@cs.rochester.edu Dept. of Computer Science Office: +1-716-275-0922; Lab: +1-716-275-1083 University of Rochester Fax: +1-716-461-2018 Rochester, NY 14627-0226 http://www.cs.rochester.edu/u/ringger/ ||||| | | | | | | | | | | | | | | | | ||||| -- Eric A. Dubiel; http://www.ilstu.edu/~eadubie mailto:eadubie@rs6000.cmp.ilstu.edu Wars CREATE Gov't Bureaucracy. MIME,HTML,NeXT,PGP,SUN Mail ok "Medicine will cure death and government will repeal taxes before Steve will fail." -Guy Kawasaki (Fall 1991 NeXTWORLD Magazine) VIEWS EXPRESSED ARE ENTIRELY MY OWN
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <5ud2i7$np2$8434@kayrad.ziplink.net> Control: cancel <5ud2i7$np2$8434@kayrad.ziplink.net> Date: 01 Sep 1997 08:07:56 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5ud2i7$np2$8434@kayrad.ziplink.net> Sender: Cheri@aol.com Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Newsgroups: comp.sys.next.programmer Subject: Re: Comparing NS Development Time Date: 1 Sep 1997 09:31:23 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5ue21b$6oe$1@concorde.ctp.com> References: <lukas-3008972357410001@lanrower2-3.urz.unibas.ch> lukas@iso.iso.unibas.ch (Lukas Schuepbach) wrote: >I am thinking about starting a new software project. I have already made >several projects with METROWERKS C++ and I would like to know what >developement time I have to expect compared to the Mac OS when I am >developing for NS. > I guess any estimate is wrong by definition since it highly depends on many factor : - How familiar you are with the tools and the frameworks (= don't waste time doing unnecessary work) - How much functionality of your application can be handled by already existing ObjectWare (the code that you don't have to write...it really can speed up things dramatically. EOF, WOF, MiscKit, etc...) As with every OO system you should plan for the long term...means spend some "useless" time during the first project to create classes/categories which really can be reused. Aloha Tomi P.S: One of the major benefits of ObjC compared to C++ are categories, since they allow you to create many useful convenience methods for classes which are part of existing frameworks without forcing you to do subclassing. This results in very reusable code. The rule of thumb is...if a category can do the job...don't subclass.
Date: Mon, 01 Sep 1997 07:52:32 -0400 From: joe.ragosta@dol.net (Joe Ragosta) Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.mac.advocacy Subject: Re: "The OpenStep Specification" Message-ID: <joe.ragosta-ya02408000R0109970752320001@news.dol.net> References: <340A6D61.6A88@rs6000.cmp.ilstu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit In article <340A6D61.6A88@rs6000.cmp.ilstu.edu>, "Eric A. Dubiel" <eadubie@rs6000.cmp.ilstu.edu> wrote: > The following message was posted on the GNUstep mailing list not so long > ago, and there was never a satisfactory answer, can anyone comment on > this who has some authoritative knowledge. Perhaps "Apple Computer, > NeXT Campus" employees?! ;-) > > Thanks! > -Eric > > What is to become of "The OpenStep Specification", now that Apple is > apparently modifying and expanding it in the process of building > Rhapsody? Will all components of the Yellow Box be considered part of > the "new and improved" OpenStep? Does Sun still have any say in its > evolution? Has it been submitted to a standards body, as was once > promised? If not, are there still plans to do so? Having learned my lesson (on DPS for Yellow Box), I'm not going to comment on OpenStep / Yellow Box issues. However, I can recommend looking at the Rhapsody FAQ on Apple's site. It's fairly complete and is kept up to date. If there is an official position on your questions, it's probably on that site. -- Regards, Joe Ragosta joe.ragosta@dol.net Visit the Complete Macintosh Web Site http://www.dol.net/~Ragosta/complmac.htm
From: penrose@w09.sfc.keio.ac.jp Newsgroups: comp.sys.next.programmer Subject: Where are art thou POSIX termios functions? Date: 01 Sep 1997 19:39:54 +0900 Organization: Keio University Shonan Fujisawa Campus, Fujisawa Japan Message-ID: <wzzppxl3lx.fsf@w09.sfc.keio.ac.jp> Howdy Folks! I am running Openstep 4.1 on an intel machine. I can't find these POSIX functions anywhere: /bin/ld: Undefined symbols: _cfsetispeed _cfsetospeed _tcgetattr _tcsetattr Under NeXTstep 3.2 these functions were available in -lposix (/usr/lib/libposix.a). Where have they gone? Is 4.1 less POSIX compliant than 3.2? I hope not. However, I have heard mention of NSSerialPort. Does such a class exist? In 4.2? In Rhapsody? Hidden in 4.1? Thanks for any answers! Christopher Penrose penrose@sfc.keio.ac.jp
Date: Mon, 01 Sep 1997 12:22:01 -0600 From: David Stes <stes@can.nl> Subject: Re: Objective-C Fragile Base Class Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Message-ID: <873133893.28718@dejanews.com> Organization: Computer Algebra Netherlands References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <EFqDnH.I15@basil.icce.rug.nl> In article <EFqDnH.I15@basil.icce.rug.nl>, tom@basil.icce.rug.dev.null.nl (Tom Hageman -- remove .dev.null to reply) wrote: > Blocks are quite cool, I agree. Does it work, like in SmallTalk's do:, on > each object in a collection of objects? (Forgive my ignorance, I haven't had > a chance to look at POC yet.) Blocks are used for an approach to exception handling, where you can specify a *default* handler at the same location where you raise an exception. The default handler can then (to be informative) refer to locals (e.g. an index that was out of bounds) of the enclosing method. The Block object, that receives a -raise message, is the default handler. The default handler, if the exception were "uncaught", does long-jump out of the method only *after* evaluating the default handler Block. Code that wants to handle an exception, uses the method -ifException:. The receiver of -ifException: is a Block that might raise an exception. The argument of -ifException: is an exception handler (also a Block). David. -------------------==== Posted via Deja News ====----------------------- http://www.dejanews.com/ Search, Read, Post to Usenet
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Objective-C Fragile Base Class Date: 01 Sep 1997 21:05:48 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x73enoetwz.fsf@tom.ics.ele.tue.nl> References: <01bcb30b$32787ae0$0100000a@ivanova> <5u1piq$khh@shelob.afs.com> In-reply-to: Greg_Anderson@afs.com's message of 27 Aug 1997 17:53:30 GMT cc: Greg_Anderson@afs.com In article <5u1piq$khh@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: You'd need to get rid of the nesting concept, such that each class had its own ivar template in the runtime system. Another approach is to associate for each class an offset with each ivar, or, like it is done in TOM, with each unit of ivars, i.e. class or extension (which is a category that may introduce new instance variables). The overhead is minimal, and you can add new ivars _at the end_ without needing any recompilation anywhere. My bigger complaint is the ongoing lack of class variables. Now there's a case where the workarounds are really messy, and the fix would not be difficult to add to the runtime structures. I'd be interested in a description of this case. TOM provides three kinds of class variables: `normal' (the one you seem to be wanting really bad), static (very similar to static member variables in C++ or, without the scoping, static variables in C), and local to a thread (which are also static in that subclasses do not see a different value). In the (quick count) 179 classes in the two standard TOM libraries, a considerable number of which aren't real classes but protocol-like behavioural classes, there are some 65 static class variables, 4 variables local to the current thread (thread id, thread object, current runloop and a number conversion buffer used curing printing), and only 3 normal class variables: the isa and asi introduced by the TOM equivalent of Objective-C's Object class (the asi is used for object administration), and a third case which appears to be an error and which should be static as well. So, in _most_ cases static variables suffice, but obviously true class variables are useful when you need them. Not that these statistics on the standard libraries justify their existence :-) More information on TOM is at http://tom.ics.ele.tue.nl:8080/. --Tiggr
From: scholz@leo.org (Bernhard Scholz) Newsgroups: comp.sys.next.programmer Subject: Re: EOF2.0 Examples not working with FlatFile? HELP! Date: 1 Sep 1997 23:52:17 GMT Organization: [posted via] Leibniz-Rechenzentrum, Muenchen (Germany) Distribution: world Message-ID: <5ufkfh$rvi$1@xeres.hsh.stusta.mhn.de> References: <5ufj4s$rih$2@xeres.hsh.stusta.mhn.de> scholz@leo.org (Bernhard Scholz) wrote: >I'm just wondering whether the EOF 2.0 example are working with >the FlatFile DB Adapter. > It works. After trying severl different compilation modes I figured out that you have to: - give 'configure_models' and absolute path for the FlatFile database - rebuild and reinstall the FlatFile adapter after running make (I don't know why the make file breaks the adapter!) BTW: EOF with FlatFile adapter looks like a damned good piece of work! Greetings, Bernhard. -- Bernhard Scholz http://www.leo.org/~scholz/ Peanuts FTP Admin http://peanuts.leo.org/ scholz@leo.org, (StuSta ONLY: boerny@xenia.hsh.stusta.mhn.de)
From: longsine_nospam_@platinum.com (Gary W. Longsine) Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.mac.advocacy Subject: Re: "The OpenStep Specification" Date: 2 Sep 1997 04:56:54 GMT Organization: Save the Skeet Foundation Message-ID: <5ug6am$rlm$1@news.platinum.com> References: <340A6D61.6A88@rs6000.cmp.ilstu.edu> Cc: eadubie@rs6000.cmp.ilstu.edu In <340A6D61.6A88@rs6000.cmp.ilstu.edu> "Eric A. Dubiel" wrote: > What is to become of "The OpenStep Specification", now that Apple is > apparently modifying and expanding it in the process of building > Rhapsody? Will all components of the Yellow Box be considered part of > the "new and improved" OpenStep? Does Sun still have any say in its > evolution? Has it been submitted to a standards body, as was once > promised? If not, are there still plans to do so? The OpenStep specification is probably not foremost in the minds of Apple these days. It seems likely that some elements of the Rhapsody enhancements to OPENSTEP will be folded into an OpenStep 2.0 spec, if there is an interested party in the future. Such a party would be another OS vendor, IBM, SGI, maybe DEC, SCO, or HP. There aren't as many as there used to be. Don't hold your breath. In any case, many of the changes could be considered proprietary extensions to OpenStep, so long as Rhapsody remains capable of running apps that were developed according to the spec. /gary - -- --- ----- ------- ----------- ------- ----- --- -- - Gary W. Longsine | This Space For Rent.
From: szallies@energotec.de (Constantin Szallies) Newsgroups: comp.sys.next.programmer Subject: Re: Objective-C Fragile Base Class Date: 2 Sep 1997 06:35:27 GMT Organization: Technet GmbH Message-ID: <5ugc3f$2gv$1@oxygen.technet.net> References: <01bcb30b$32787ae0$0100000a@ivanova> "Robert Fisher" <rfisher@onr.com> wrote: >Apple's latest Rhapsody FAQ reveals that Objective-C does have something of >a fragile base class problem. To sum up: Adding instance variables to a >class breaks code that depends on that class. (The "broken" code would need >to be recompiled.) The "fragile base class problem" is the observation that you may have to reorganize your inheritance hierarchy when you do changes in your class library. It's not about interface dependencies. Greetings -- # Constantin Szallies, Energotec GmbH # szallies@energotec.de # http://www.energotec.de/~szallies/ # 49211-9144018
Newsgroups: comp.sys.next.programmer Subject: Re: Where are art thou POSIX termios functions? References: <wzzppxl3lx.fsf@w09.sfc.keio.ac.jp> From: frank@ifi.unibas.ch Message-ID: <340bc356.0@maser.urz.unibas.ch> Date: 2 Sep 97 07:42:14 GMT penrose@w09.sfc.keio.ac.jp wrote: > > Howdy Folks! > > I am running Openstep 4.1 on an intel machine. I can't find these POSIX > functions anywhere: > > /bin/ld: Undefined symbols: > _cfsetispeed > _cfsetospeed > _tcgetattr > _tcsetattr > > Under NeXTstep 3.2 these functions were available in -lposix > (/usr/lib/libposix.a). > > Where have they gone? Is 4.1 less POSIX compliant than 3.2? I hope not. They have gone to nowhere! !#@$&#@$ NeXT dropped posix in OS4.x. You will have to compile under 3.3 and then use that binary, as the 3.3 compatibility still works fine. By the way, even those posix functions which are still around are broken, broken, broken. -Robert -- Institut fuer Informatik tel +41 (0)61 321 99 67 Universitaet Basel fax. +41 (0)61 321 99 15 Robert Frank Mittlere Strasse 142 rfc822: frank@ifi.unibas.ch (NeXT,MIME mail ok) CH-4056 Basel X400: S=frank;OU=ifi;O=unibas;P=switch;A=arcom;C=ch Switzerland
Newsgroups: comp.sys.next.programmer Subject: samba & posix & NEXTSTEP/OPENSTEP From: frank@ifi.unibas.ch Message-ID: <340bd348.0@maser.urz.unibas.ch> Date: 2 Sep 97 08:50:16 GMT I'm trying to compile the samba version 1.9.17alpha4 for NEXTSTEP/OPENSTEP. I've given up trying OS4.2, as the new version uses some posix/sysv functions which would have to be mapped to some bsd functions. So I resorted to use NS3.3 and the posix flag. Compilation is fine (had to tune the include.h file to use posix includes) but execution isn't. Just trying to run the smbpasswd program fails with the message that the password file could not be locked due to an invalid argument. I isolated the procedure and tried a simple test program - no luck. Looks like the fnctl is broken and won't even accept the flock structure! The routine: int p_lock(int fd, int type) { struct flock lock; lock.l_type = type; /* F_WRLCK or F_UNLCK */ lock.l_whence = 0; /* SEEK_SET */ lock.l_start = 0; /* from the first byte */ lock.l_len = 0; /* the whole file */ lock.l_pid = 0; return fcntl(fd, F_SETLK, &lock); } returns -1 and errno is set to EINVAL (invalid argument)! According to the man pages, the data is correct. Anyone had the same problems? Do I *have* to use the flock function instead? -Robert -- Institut fuer Informatik tel +41 (0)61 321 99 67 Universitaet Basel fax. +41 (0)61 321 99 15 Robert Frank Mittlere Strasse 142 rfc822: frank@ifi.unibas.ch (NeXT,MIME mail ok) CH-4056 Basel X400: S=frank;OU=ifi;O=unibas;P=switch;A=arcom;C=ch Switzerland
From: marcel@system.de Newsgroups: comp.sys.next.programmer Subject: Re: Objective-C Fragile Base Class Date: 2 Sep 1997 09:21:22 GMT Organization: Technical University Berlin, Germany Distribution: world Message-ID: <5uglqi$4qe$1@brachio.zrz.TU-Berlin.DE> References: <x73enoetwz.fsf@tom.ics.ele.tue.nl> In article <x73enoetwz.fsf@tom.ics.ele.tue.nl> tiggr@ics.ele.tue.nl (Pieter Schoenmakers) writes: > In article <5u1piq$khh@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > > You'd need to get rid of the nesting concept, such that each class had > its own ivar template in the runtime system. > > Another approach is to associate for each class an offset with each ivar, > or, like it is done in TOM, with each unit of ivars, i.e. class or > extension (which is a category that may introduce new instance variables). > The overhead is minimal, and you can add new ivars _at the end_ without > needing any recompilation anywhere. Objective-C does provide a mechanism that could be made to work this way: indexed instance variables. Alas, many of the Foundation classes override +alloc, so it would be hard to sneak them in. There shouldn't be too much of a problem with user-provided classes. Marcel
From: "Lee Byeong-ho" <bhlee@cnt.co.kr> Newsgroups: comp.sys.next.programmer Subject: Java Programming in OpenStep4.2 Date: 2 Sep 1997 09:27:09 GMT Organization: HanQ.Net Message-ID: <01bcb782$4befa7c0$68a77fd2@dooly> References: <beauvois-ya023580000908970504060001@news.utk.edu> <5snruk$kom@queens.netuse.de> Hello, All... I'm using OpenStep Mach 4.2 for Intel and I want to program with java. So I download JDK1.1 for free BSD and ... But I couldn't execute any commands such as javac, jar, ... Please give me tips on Java programming with OpenStep. Best Regards..... /Byeong-ho, Lee
From: "Robert Fisher" <rfisher@onr.com> Newsgroups: comp.sys.next.programmer Subject: Is Java Better than Objective-C? Date: 2 Sep 1997 14:51:32 GMT Organization: Renegade Software Message-ID: <01bcb4cf$9b253c70$0100000a@ivanova> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> Thanks to everyone who responded to my "Objective-C's FBC" post. It cleared some things up for me and confirmed other thoughts I had. I wrote, regarding Objective-C's fragile base class problem (OK, "problem" may be to strong a word)... > Why shouldn't > Apple fix this? Tom Hageman -- remove .dev.null to reply <tom@basil.icce.rug.dev.null.nl> wrote in article <EFL8FL.7oE@basil.icce.rug.nl>... > They do. They want you to use Java ;-) Thanks for the segue. =) So, does Java have a fragile base class problem? There has been lots of discussion about the differences between Java and C++, but I don't think I've seen much about the differences between Java and Objective-C. Can anyone offer a list of Objective-C features that an Objective-C programmer would miss if he were forced to use Java? For instance, Java doesn't have anything akin to Objective-C's ForwardInvocation method, does it? -- Robert Fisher Renegade Software rfisher@onr.com
From: nurban@sps1.phys.vt.edu (Nathan Urban) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 2 Sep 1997 12:11:22 -0400 Organization: Data Systems Consulting, Inc. Message-ID: <5uhdra$qre$1@sps1.phys.vt.edu> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> In article <01bcb4cf$9b253c70$0100000a@ivanova>, "Robert Fisher" <rfisher@onr.com> wrote: > Can anyone offer a list of Objective-C features that an Objective-C > programmer would miss if he were forced to use Java? Here is a partial list, but it focuses more on their similarities (which are greater than one might first imagine): http://www.misckit.com/DynamicOO/javaobjccomp.html
From: nospam+next@luomat.peak.org (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer Subject: Re: Java Programming in OpenStep4.2 Date: 2 Sep 1997 17:28:52 GMT Organization: http://www.peak.org/~luomat/ Message-ID: <5uhick$4bf$1@ha1.rdc1.nj.home.com> References: <beauvois-ya023580000908970504060001@news.utk.edu> <5snruk$kom@queens.netuse.de> <01bcb782$4befa7c0$68a77fd2@dooly> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: bhlee@cnt.co.kr In <01bcb782$4befa7c0$68a77fd2@dooly> "Lee Byeong-ho" wrote: > Please give me tips on Java programming with OpenStep. There is only one tip: ``It can't be done, wait for Rhapsody'' TjL -- "NeXT continually embarrassed the rest of the industry for having workable, shipping, and high-quality answers before most people knew the questions." -- Gary Longsine, full text at: http://www.stepwise.com/Articles/Editorial/TheHolyGrail.html
From: hhoff@ragnarok.en.REMOVETHIS.eunet.de (Holger Hoffstaette) Newsgroups: comp.sys.next.programmer Subject: Re: samba & posix & NEXTSTEP/OPENSTEP Date: 2 Sep 1997 19:15:03 GMT Organization: N.eXTers W.ith A.ttidude Message-ID: <5uhojn$g2@ragnarok.en.eunet.de> References: <340bd348.0@maser.urz.unibas.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit frank@ifi.unibas.ch wrote: > I'm trying to compile the samba version 1.9.17alpha4 for > NEXTSTEP/OPENSTEP. Use the release version of 1.9.17! > I've given up trying OS4.2, as the new version uses some posix/sysv > functions which would have to be mapped to some bsd functions. The release version compiles fine on 4.2. Just uncomment the platform in the Makefile and go! No -posix brain drain required. Holger -- holger"at"ragnarok.en.eunet.de (NeXTMail, MIME) 'Where I come from it's important to develop stuff that doesn't suck.' -- David Young <dwy"at"ace.net>
From: robert@next.amistep.osn.de (Robert Hofmann) Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,de.comp.sys.next Subject: S: PGP for NS 3.3 Intel Date: 1 Sep 1997 19:26:41 GMT Organization: OSN, Onlineservice Nuernberg, Germany Message-ID: <5uf4th$ijo$1@news.osn.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Hi, I've just tried to compile pgp 2.6.3, but it seems that an include for strdup() is missing on my system. Well, I am not a very good c-coder, so maybe someone can send me a full-working (compiled) version of pgp for NeXTSTEP-Intel? Please reply via email only! Tnx! -- _ o Robert Hofmann | robert@next.amistep.osn.de, rhofmann@lucent.com |<)_/# | http://www.osn.de/user/amistep TT <T Amiga & NeXTSTEP | 2:2490/1015@fidonet, 39:171/101@amiganet
From: Steve Dekorte <dekorte@slip.net> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 2 Sep 1997 20:40:11 GMT Organization: Slip.Net (http://www.slip.net) Message-ID: <5uhtjb$35s$1@owl.slip.net> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <5ua3k8$nmt$1@owl.slip.net> <5uab67$fgp$1@nntp2.ba.best.com> In comp.lang.objective-c Kenneth C. Dyke <kcd@babylon5.jumpgate.com> wrote: > On 08/30/97, Steve Dekorte wrote: > >As long as the subclasses use accessor methods for all variable > references, > >there should be no need to recompile even subclasses. > Actually, if a subclass has it's own instance variables, you will > run the risk of accidentally overwriting your superclass's instance > variables if some got added and you don't recompile. Likewise, > the superclass would potentially be overwriting your instance variables > as well. That's right. I'm so used to using variable lists that I didn't think about how most people still use ivars in subclasses. They will have to recompile. (and won't have the freedom to add/remove variables at runtime or have to archiving and copying done in a clean way) -- Steve Dekorte - OpenStep consultant - San Francisco
From: alex@guava.phil.lehigh.edu (Alex Levine) Newsgroups: comp.sys.next.programmer Subject: Anyone know how to implement substring search in WebScript? Date: 2 Sep 1997 14:03:10 GMT Distribution: world Message-ID: <5uh6au$1q3o@fidoii.cc.Lehigh.EDU> Subject says it all. Infinite gratitude for all replies. Alex -- Alexander Levine Philosophy Department Lehigh University ATL2@lehigh.edu
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Objective-C Fragile Base Class Date: 03 Sep 1997 01:09:59 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x790xfmhx4.fsf@tom.ics.ele.tue.nl> References: <x73enoetwz.fsf@tom.ics.ele.tue.nl> <5uglqi$4qe$1@brachio.zrz.TU-Berlin.DE> In-reply-to: marcel@system.de's message of 2 Sep 1997 09:21:22 GMT In article <5uglqi$4qe$1@brachio.zrz.TU-Berlin.DE> marcel@system.de writes: Objective-C does provide a mechanism that could be made to work this way: indexed instance variables. If something doesn't work that way but can be made to do so, it denotes a very bad approach, since, usually, it still won't work that way. If the language or tools do not provide something automatically, it ends up not being used. Alas, many of the Foundation classes override +alloc, so it would be hard to sneak them in. There shouldn't be too much of a problem with user-provided classes. But you'd still have the same problem with *any* library from another source than yourself. That's not usability; that's a pain. TOM is at http://tom.ics.ele.tue.nl:8080/. --Tiggr
From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Newsgroups: comp.sys.next.programmer Subject: NSColor encoding for user defaults really missing ? Date: 2 Sep 1997 23:21:13 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5ui719$i91$1@concorde.ctp.com> Hi, I just spend a couple of hours working on a MiscKit extension to easiely store NSColors in the user defaults... ... however...the more I coded the less I could believe that NeXT did not already provide a way to encode and decode colors. While there is a set of methods for dealing with colors on the pasteboard...I was not able to find something inside the docu....but since the AppKit fucntion docu is not avialable for 4.x I slowly think that I propably wasted a lot of time on that on... Could somebody slap me against the head by giving me a pointer to a NeXT supplied function...or make me feel good by confirming that this functionaliy is really missing ? Thanx. Aloha Tomi
From: Jochen Klinke <jklinke@ucsd.edu> Newsgroups: comp.sys.next.programmer Subject: compiling gas-2.2 under NSFIP3.3 Date: Tue, 02 Sep 1997 16:27:11 -0700 Organization: Scripps Institution of Oceanography Message-ID: <340CA0CF.D4D9E9F5@ucsd.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I'm trying to compile the GNU Assembler (gas-2.2) on a NSFIP3.3 host for a m68k target system. During the make of the file 'trad-core.c' the compiler exits with the message: /NextDeveloper/Headers/bsd/sys/user.h:225 illegal declaration, missing ';' after 'u_kproc' /NextDeveloper/Headers/bsd/sys/user.h:225 illegal method selector, found '1' /NextDeveloper/Headers/bsd/sys/user.h:225 illegal method definition, missing '{' after ']' /NextDeveloper/Headers/bsd/sys/user.h:333 illegal external declaration, found '}' Any ideas on what I'm doing wrong or how to fix this? Thanks, jk jklinke@ucsd.edu
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 03 Sep 1997 01:17:07 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x77mczmhl8.fsf@tom.ics.ele.tue.nl> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> <5uhdra$qre$1@sps1.phys.vt.edu> In-reply-to: nurban@sps1.phys.vt.edu's message of 2 Sep 1997 12:11:22 -0400 In article <5uhdra$qre$1@sps1.phys.vt.edu> nurban@sps1.phys.vt.edu (Nathan Urban) writes: Here is a partial list... The comp.lang.Objective-C FAQ has the following entry: The most obvious difference, apart from all the applet hype, is syntax: Java's syntax is based on C++ syntax whereas Objective-C employs the C syntax (and semantics) with object declaration and manipulation based on Smalltalk. The most basic difference between Java and Objective-C is that they employ exactly the same object model. (Nested classes do not change that since they are retrofitted on the original Java 1.0 object model for compatibility with 1.0 tools.) In fact, Java objects are much more Objective-C with a different syntax, than they are C++. Other differences include mandatory typing, more extensive thread support, security managers, name spaces, predefined classes (java.lang, the AWT, etc), etc. to which I should add that, as the original poster hinted with his remark on Invocation objects, Java provides no selector type, and without a selector type you can not provide the power of dynamic binding to the programmer, since [my_array makeElementsPerform: @selector (foo)] becomes impossible. Sure, things like that can be retrofitted on top of Java's introspection, but I doubt its speed, and it still won't be part of the language, thus proprietary or peculiar, thus unavailable to the general programmer and thus unused. Something that is not part of the language tends not to be used. Otherwise, C would be object oriented. --Tiggr
From: lhow@ecr.mu.oz.au (Luke HOWARD) Newsgroups: comp.sys.next.programmer Subject: Re: Java Programming in OpenStep4.2 Date: 2 Sep 1997 23:59:43 GMT Organization: Comp Sci, University of Melbourne Message-ID: <5ui99f$b6e@mulga.cs.mu.OZ.AU> References: <beauvois-ya023580000908970504060001@news.utk.edu> <5snruk$kom@queens.netuse.de> <5uhick$4bf$1@ha1.rdc1.nj.home.com> Timothy J. Luoma (nospam+next@luomat.peak.org) wrote: : In <01bcb782$4befa7c0$68a77fd2@dooly> "Lee Byeong-ho" wrote: : > Please give me tips on Java programming with OpenStep. : There is only one tip: : ``It can't be done, wait for Rhapsody'' Not entirely true. Earlier versions of the JDK worked (without the AWT) were ported to NEXTSTEP and worked fine, at least for my use. -- Luke I don't knwo from the top of my head...but I think that in this case stringValue really returns a "nil" which in this case is a value you don't expect since you can't set it...) But: [myColorWell setColor:nil] works like a charm which IMHO makes absolutly no sense since the color well will display a colorBackgroundGray "color" (or some random color) and lure you to believe that the well contains some useful information. It also acts a little unpredictalble...since sometimes you get some funky color (say Pantone blue)...but when you drag the swatch it is a black swatch. Personally I would switch that behavior since nil for textFields makes sense..but definitly not for colorWells. And while we are at it...it gets even better: (gdb) po [desktopColorPrefWell setColor:nil] <NSColorPanel: 0x2c66cc> Hmm... shouldn't that method return void ? Why is GDB actually returning a valid object and not barfing at me as it usually does. Let's not call it a feature but a bug. Comments ? Aloha Tomi
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <5uib9l$3s7@examiner.concentric.net> Control: cancel <5uib9l$3s7@examiner.concentric.net> Date: 03 Sep 1997 01:11:39 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5uib9l$3s7@examiner.concentric.net> Sender: FREE CABLE<pifer@laugh.com> Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: jesjones@halcyon.com (Jesse Jones) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: Tue, 02 Sep 1997 21:31:29 -0700 Organization: Edmark Message-ID: <jesjones-ya02408000R0209972131290001@news.halcyon.com> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit In article <872781306.21356@dejanews.com>, David Stes <stes@can.nl> wrote: > In article <x7yb5mh60k.fsf@tom.ics.ele.tue.nl>, > tiggr@ics.ele.tue.nl (Pieter Schoenmakers) wrote: > [snip] > > Using structs and unions obviously is an error and the use by NeXT > > of NSPoint, NSRect and NSSize is something I don't fancy, not even when > > the argument in favour is `efficiency'. > > The Point and Rectangle classes of ICpak101 are nice. > > In ICpak201 (for X/Windows) you have then nifty methods > like -moveTo:aPoint or -sizeTo:aPoint. But effiency is important, especially for something as commonly used as points and rectangles. It may be heresy, but C++ excels at these sort of low level classes: operator overloading allows you to use a very natural syntax and because they can be stack based and C++ supports inlining they're as efficient as raw structs. Is it possible that some simple C++ classes will be added to the AppKit? Or are there a lot of people not using Objective-C++? --Jesse
From: mtrombin@ix.netcom.com (Mark Trombino) Newsgroups: comp.sys.next.programmer,comp.sys.next.bugs Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 3 Sep 1997 07:01:38 GMT Organization: Egghead Billy, Inc. Message-ID: <5uj20i$cr9@sjx-ixn9.ix.netcom.com> References: <5ui9k7$io4$2@concorde.ctp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii In <5ui9k7$io4$2@concorde.ctp.com> Thomas Engel wrote: > is it just me or are there other people who consider this as a bug: > > [myTextField setStringValue:nil] > > results in an exception...but for a textField I think it might make sense to > resert the content to the state of a freshly created textField. (I don't knwo > from the top of my head...but I think that in this case stringValue really > returns a "nil" which in this case is a value you don't expect since you > can't set it...) > > But: > > [myColorWell setColor:nil] > > works like a charm which IMHO makes absolutly no sense since the color well > will display a colorBackgroundGray "color" (or some random color) and lure > you to believe that the well contains some useful information. > It also acts a little unpredictalble...since sometimes you get some funky > color (say Pantone blue)...but when you drag the swatch it is a black swatch. I think that [myTextField setStringValue:@""] makes more sense than [myTextField setStringValue:nil] but it would be nice if setting a string value to nil had the same effect. I agree with you about the color wells though.... -- |-------------------------------------------------------------------------| | Mark Trombino | J A M S o f t | | mtrombin@ix.netcom.com | Audio DSP Tools for Openstep & Rhapsody | |-------------------------------------------------------------------------|
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <5uj3o4$4er@news.jumpnet.com> Control: cancel <5uj3o4$4er@news.jumpnet.com> Date: 03 Sep 1997 07:31:22 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5uj3o4$4er@news.jumpnet.com> Sender: Br. Michael<stanthony@jumpnet.com> Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 03 Sep 1997 09:08:36 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7afhuyivf.fsf@tom.ics.ele.tue.nl> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> In-reply-to: jesjones@halcyon.com's message of Tue, 02 Sep 1997 21:31:29 -0700 In article <jesjones-ya02408000R0209972131290001@news.halcyon.com> jesjones@halcyon.com (Jesse Jones) writes: Is it possible that some simple C++ classes will be added to the AppKit? NSPoint et al already are what you desire; a struct is after all nothing but a class with all members public. --Tiggr
From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Newsgroups: comp.sys.next.programmer Subject: Automatic spinning cursor of Mach not going away ?? Date: 3 Sep 1997 10:31:24 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5uje9s$1q2$1@concorde.ctp.com> Hi, NeXTs automatic wait cursor (for which they even own a patent!) is quite cool and usually just does the right thing. But for one of my applications, under certain yet unknown circumstances, it starts spinning...and never stops. The application is fully active and works just fine..but the cursor is spinning and I have not found an obvious way to reset it. [[NSCursor arrowCursor] set] does not cure the disease :-( Is there any hidden DPS code that could be used or some other evil low level trick ? What is the internal mechanims that triggers teh spinning cursor..maybe there is a way to find out what causes it to spinn ( I suspect that maybe some of the NSWorkspace requests are part of the problem..but couldn't nail them down yet) Any ideas ? Aloha Tomi
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <5ujkl7$jc9$2097@news.freedomnet.com> Control: cancel <5ujkl7$jc9$2097@news.freedomnet.com> Date: 03 Sep 1997 12:42:18 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5ujkl7$jc9$2097@news.freedomnet.com> Sender: rgr@freedomnet.com Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: eharley@pacbell.net (Eric Harley) Newsgroups: comp.sys.next.programmer Subject: Re: Java Programming in OpenStep4.2 Date: Wed, 03 Sep 1997 06:49:41 -0800 Organization: Harley Consulting Message-ID: <eharley-0309970649420001@ppp-207-214-149-17.snrf01.pacbell.net> References: <beauvois-ya023580000908970504060001@news.utk.edu> <5snruk$kom@queens.netuse.de> <5uhick$4bf$1@ha1.rdc1.nj.home.com> <5ui99f$b6e@mulga.cs.mu.OZ.AU> In article <5ui99f$b6e@mulga.cs.mu.OZ.AU>, lhow@ecr.mu.oz.au (Luke HOWARD) wrote: > Timothy J. Luoma (nospam+next@luomat.peak.org) wrote: > > : In <01bcb782$4befa7c0$68a77fd2@dooly> "Lee Byeong-ho" wrote: > > : > Please give me tips on Java programming with OpenStep. > > : There is only one tip: > > : ``It can't be done, wait for Rhapsody'' > > Not entirely true. Earlier versions of the JDK worked (without > the AWT) were ported to NEXTSTEP and worked fine, at least > for my use. Whatever happened to blackjava? did the support for that end or something? -- Eric Harley --------------------------------------------------------------------- "Medicine will cure death and government will repeal taxes before Steve will fail." -Guy Kawasaki (Fall 1991)
Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,de.comp.sys.next From: G.C.Th.Wierda@AWT.nl (Gerben Wierda) Subject: Re: S: PGP for NS 3.3 Intel Message-ID: <EFxp0o.2FG@AWT.NL> Sender: news@AWT.NL Organization: Adviesraad voor het Wetenschaps- en Technologiebeleid References: <5uf4th$ijo$1@news.osn.de> Date: Wed, 3 Sep 1997 13:19:36 GMT robert@next.amistep.osn.de (Robert Hofmann) wrote: > >Hi, > >I've just tried to compile pgp 2.6.3, but it seems that an include for >strdup() is missing on my system. > >Well, I am not a very good c-coder, so maybe someone can send me a >full-working (compiled) version of pgp for NeXTSTEP-Intel? Why not just download it from: ftp://ftp.nluug.nl/pub/comp/next/PGP/ Yours, -- Gerben Wierda, Stafmedewerker Adviesraad voor het Wetenschaps- en Technologiebeleid. Staff member Advisory Council for Science and Technology Policy Javastraat 42, 2585 AP, Den Haag/The Hague, The Netherlands Tel (+31) 70 3639922 Fax (+31) 70 3608992 http://www.AWT.nl/prive/wierda/ The fox knows many things, but the hedgehog knows one big thing.
From: Konstantin Wiesel <kwiesel@mailhost.jura.uni-bonn.de> Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,de.comp.sys.next Subject: Re: S: PGP for NS 3.3 Intel Date: Wed, 3 Sep 1997 16:18:13 +0000 Organization: RHRZ - University of Bonn (Germany) Message-ID: <Pine.NXT.3.95.970903161349.4695D-100000@ikarus.jura.uni-bonn.de> References: <5uf4th$ijo$1@news.osn.de> <EFxp0o.2FG@AWT.NL> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII In-Reply-To: <EFxp0o.2FG@AWT.NL> On Wed, 3 Sep 1997, Gerben Wierda wrote: > robert@next.amistep.osn.de (Robert Hofmann) wrote: > > > >Hi, > > > >I've just tried to compile pgp 2.6.3, but it seems that an include for > >strdup() is missing on my system. > > > >Well, I am not a very good c-coder, so maybe someone can send me a > >full-working (compiled) version of pgp for NeXTSTEP-Intel? > Das Thema gab es schon einmal hier, damals loeste jemand das Problem indem er "-d strdup=NXCopyBuffer" als zusaetzlichen compile Parameter angab. ueber Die Definition habe ich jetzt aus dem Gedaechtnis, das sollte man also in der Developer Doku nochmal nachschauen, jedenfalls hat das bei mir ebenfalls geklappt. Regards Konstantin Wiesel Email:kwiesel@mailhost.jura.uni-bonn.de
From: alkfjsadlj@ljfljsd.com Subject: Pamela & Tommy Lee's Secret Video Newsgroups: comp.sys.next.programmer Message-ID: <33fb85d7.0@news.inreach.com> Date: 21 Aug 97 00:03:35 GMT Watch Babewatch Star Pamela Anderson Lee and Husband Tommy Lee's Secret Sex tape that is from there Malibu House. This tape is the only orignal copyrighten tape in existest that shows Pamela and Tommy Bare all. Don't be fooled by others, this the one and only orignal tape around. Visit: http://www.pamsex.com
From: Jim Redman <jim@ergotech.com> Newsgroups: comp.sys.next.programmer Subject: Stopping NSRunLoops Date: Tue, 02 Sep 1997 16:36:23 -0600 Organization: Southwest Cyberport Message-ID: <340C94E7.6BE2@ergotech.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I feel that this should be a fairly dumb question, but just how do you stop an NSRunLoop that has been started with a -run? Is it just in the documentation and I missed it, or is there something clever about this? My major interest is stopping the loop on a thread so that the thread can exit cleanly and clear up the NSAutoReleasePool. Just for the record, this application does not use AppKit, just Foundation. Jim
From: ga@ed4u.com (G. Apple) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 3 Sep 1997 17:11:15 GMT Organization: Advanced Communications Engineering, Inc. Message-ID: <5uk5nj$8vh$2@gte1.gte.net> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <x7afhuyivf.fsf@tom.ics.ele.tue.nl> In article <x7afhuyivf.fsf@tom.ics.ele.tue.nl>, tiggr@ics.ele.tue.nl (Pieter Schoenmakers) wrote: > In article <jesjones-ya02408000R0209972131290001@news.halcyon.com> jesjones@halcyon.com (Jesse Jones) writes: > > Is it possible that some simple C++ classes will be added to the > AppKit? > > NSPoint et al already are what you desire; a struct is after all nothing > but a class with all members public. --Tiggr Have you looked at Apple Class Suites? ACS contains pure C++ classes for an extensive set of functionality. Although it had roots in MacApp, it is now usable anywhere you can compile C++. It is available for download as part of MacApp Release 13, but does not require MacApp to use it. -- G. Gordon Apple, PhD The Ed4U Project Advanced Communications Engineering, Inc. Redondo Beach, CA ga@ed4u.com
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 3 Sep 1997 19:51:48 GMT Organization: Apple Computer, Inc. Message-ID: <5ukf4k$ika$1@news.apple.com> References: <5uj20i$cr9@sjx-ixn9.ix.netcom.com> Mark Trombino writes > > In <5ui9k7$io4$2@concorde.ctp.com> Thomas Engel wrote: > > is it just me or are there other people who consider this as a bug: > > > > [myTextField setStringValue:nil] > > > > results in an exception...but for a textField I think it might make > > sense to resert the content to the state of a freshly created > > textField. (I don't knwo from the top of my head...but I think that in > > this case stringValue really returns a "nil" which in this case is a > > value you don't expect since you can't set it...) 1. setStringValue expects an NSString argument 2. nil is not an NSString 3. Therefore [myTextField setStringValue:nil] raises an exception. This is consistent with the behavior of AppKit and Foundation in other areas. I don't know why you'd expect that setting the stringValue to nil would reset the field to its initial state. > > But: > > > > [myColorWell setColor:nil] > > > > works like a charm which IMHO makes absolutly no sense since the color > > well will display a colorBackgroundGray "color" (or some random color) > > and lure you to believe that the well contains some useful > > information. It also acts a little unpredictalble...since sometimes > > you get some funky color (say Pantone blue)...but when you drag the > > swatch it is a black swatch. Thanks for pointing this out. I'll file a bug on it. While the behavior isn't defined for a nil argument, we like to eliminate "unpredictable" behavior where we can. > I think that > > [myTextField setStringValue:@""] > > makes more sense than > > [myTextField setStringValue:nil] > > but it would be nice if setting a string value to nil had the same > effect. That's an interesting idea, but I think the current behavior of treating nil objects as invalid in most cases is more helpful for flushing out bugs during development. -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: seanlNoSpam@carmi.cs.umd.edu (Sean Luke) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 3 Sep 1997 20:44:27 GMT Organization: University of Maryland at College Park Message-ID: <5uki7b$ama$2@walter.cs.umd.edu> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> <5uhdra$qre$1@sps1.phys.vt.edu> Nathan Urban (nurban@sps1.phys.vt.edu) wrote: >In article <01bcb4cf$9b253c70$0100000a@ivanova>, "Robert Fisher" <rfisher@onr.com> wrote: > >> Can anyone offer a list of Objective-C features that an Objective-C >> programmer would miss if he were forced to use Java? > >Here is a partial list, but it focuses more on their similarities (which >are greater than one might first imagine): ...especially considering that Java was largely based semantically on Obj-C (yes, it's true! I got it from Naughton himself.) Sean Luke
From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 3 Sep 1997 21:38:21 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5uklcd$dmp$1@concorde.ctp.com> References: <5uj20i$cr9@sjx-ixn9.ix.netcom.com> <5ukf4k$ika$1@news.apple.com> MaRK_BeSSeY@Apple.Com (Mark Bessey) wrote: > >1. setStringValue expects an NSString argument >2. nil is not an NSString >3. Therefore [myTextField setStringValue:nil] raises an exception. >This is consistent with the behavior of AppKit and Foundation in other >areas. I don't know why you'd expect that setting the stringValue to nil >would reset the field to its initial state. Ok but: 1. stringValue is typed to a NSString 2. nil is not an NSString 3. still [myTextField stringValue] returns a nil until you set some data. :-) Ok..why I ran into this is simply because: [myDestField setStringValue:[mySourceField stringValue]] Bang'ed on me. And IMHO the value which is returned by an objects "getter" should be a valid parameter for the corresponding "setter". Besides that this behavior is not documented for NSCell or NSControl. > [NSColor stuff]... >Thanks for pointing this out. I'll file a bug on it. While the behavior >isn't defined for a nil argument, we like to eliminate "unpredictable" >behavior where we can. > Thanx a lot. >> I think that >> >> [myTextField setStringValue:@""] >> >> makes more sense than >> >> [myTextField setStringValue:nil] >> >> but it would be nice if setting a string value to nil had the same >> effect. > >That's an interesting idea, but I think the current behavior of treating >nil objects as invalid in most cases is more helpful for flushing out bugs >during development. > See above. The problem was that I was reading some contents of a textField...stored it in one of my ivars...and once I got back in to that window I wanted to restore the original value. Bang. If nil is not allowed then the behavior should be better documented Like it is for NSArrays addObject: "If anObject is nil, an NSInvalidArgumentException is raised"). Aloha Tomi
From: Mark Nowiasz <buckaroo@blackbox.free.de> Newsgroups: comp.sys.next.programmer Subject: Appending text to a text object? Date: 3 Sep 1997 21:05:22 GMT Organization: Blackbox - The Dark Site Message-ID: <5ukjei$9nk$1@blackbox.free.de> Hi! I've encountered a tricky problem: how can I append text to a text object *without* using [text setSel::] and [text replaceSel:]? Let me tell you why I need a solution for this.. I'm developing an IRC client for NeXT (I'm using a NeXTstation under NS 3.3) and I'm using a scrollview containing a text object for the output. For the user's input I'm using a ordinarily textfield. Now, using [text setSel::] creates two major problems: 1. Whenever I'm selecting a portion of the text (for example, to copy), the selection is of course disappearing whenever [text setSel::] is being called. 2. What is much more worse, [text setSel::] makes the text object the window's firstresponder, which leads to the deactivating of the textfield where one can enter text. [window makeFirstResponder:textfield] won't do. Now, when there are a lot of messages coming in (which result in a lot of calls of [text setSel::] it's practically impossible to use the textfield, since every new message deactivates its :-/ I've had had a look at the other two IRC clients existing for NeXT, Encircle and IRC 0.5. IRC 0.5 hasn't solved this problem, it uses another window for input, which I consider to be less then good. Now, Encircle on the other hand manages somehow to append text to it's text object *without* the major disadvantages mentioned above, and the nib files says that it's an ordinarily text object. Well.. does anyone have a clue how this can be made? Thanks in advance, Mark -- |\ _,,,---,,_ Mark Nowiasz (buckaroo@blackbox.free.de) /,.-'' -. ;-;;,_ PGP-Key available at request. |,4- ) )-,_..;\ ( '-' http://www.free.de/~buckaroo/ |IRC: Buckaroo '---''(_/--' -'\_) >>> Semper Paratus <<<
From: "Robert A. Decker" <comrade@umich.edu> Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: setting the environment variable for PERL5LIB not working Date: 3 Sep 97 19:08:31 -0400 Organization: University of Michigan ITD News Server Message-ID: <B0336640-69F0@141.214.128.36> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit nntp://news.itd.umich.edu/comp.sys.next.programmer How can I set it up so that all users have a PERL5LIB environment variable? When I set a line for login.std inside /etc to: set PERL5LIB=/usr/lib/perl5 it doesn't seem to work. thanks, rob -- <mailto: "Robert A. Decker" comrade@umich.edu> Listen to my Realaudio playlist:<http://hmrl.cancer.med.umich.edu/Rob/index.ssi> Programmer Analyst - Health Media Research Lab University of Michigan Comprehensive Cancer Center "Get A Life" quote #10: "Wow. I'm a genius too. I think. BEEP." -Chris Elliott
From: dlehn@crib.bevc.blacksburg.va.us (David I. Lehn) Newsgroups: comp.sys.next.programmer Subject: Re: NSColor encoding for user defaults really missing ? Date: 3 Sep 1997 23:11:36 GMT Organization: Virginia Tech Message-ID: <5ukqr8$8vq$1@solaris.cc.vt.edu> References: <5ui719$i91$1@concorde.ctp.com> Thomas Engel (tsengel@immd9.informatik.uni-erlangen.de) wrote: > I just spend a couple of hours working on a MiscKit extension to easiely > store NSColors in the user defaults... ... > ... however...the more I coded the less I could believe that NeXT did not > already provide a way to encode and decode colors. ... > Could somebody slap me against the head by giving me a pointer to a NeXT > supplied function...or make me feel good by confirming that this functionaliy > is really missing ? In Draw.app source ( propertyList.{hm} ): id propertyListFromNSColor(NSColor *color); NSColor *colorFromPropertyList(id plist, NSZone *zone); MiscKit could use a subclass of NSCoder to deal with property lists. Something like what GNUstep was heading for. Call it MiscPropertyListCoder and have some overloaded methods that let you can give a name to the encoded values so the plist is readable. -dave --- David I. Lehn <dlehn@vt.edu> | http://www.vt.edu:10021/D/dlehn/ Computer Engineering Graduate @ Virginia Tech in sunny Blacksburg, VA
From: Cosmo Roadkill <cosmo.roadkill%bofh.int%rauug.mil.wi.us@bofh.int> Newsgroups: comp.sys.next.programmer,uswest.test Subject: cmsg cancel <5uks38$aq46@utoepia.advtech.uswest.com> Control: cancel <5uks38$aq46@utoepia.advtech.uswest.com> Date: 04 Sep 1997 01:51:10 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5uks38$aq46@utoepia.advtech.uswest.com> Sender: NNTP Component <nntpcomp@microsoft.com> My From: line has been fudged because many test newsgroup autoresponders respond to control messages. My apologies! Please see the X-Cancelled-By: line for my proper address. Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: Steve Dekorte <dekorte@slip.net> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 4 Sep 1997 01:52:19 GMT Organization: Slip.Net (http://www.slip.net) Message-ID: <5ul48j$3dj$1@owl.slip.net> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> In comp.lang.objective-c Jesse Jones <jesjones@halcyon.com> wrote: > In article <872781306.21356@dejanews.com>, David Stes <stes@can.nl> wrote: > > The Point and Rectangle classes of ICpak101 are nice. > > > > In ICpak201 (for X/Windows) you have then nifty methods > > like -moveTo:aPoint or -sizeTo:aPoint. > But effiency is important, especially for something as commonly used as > points and rectangles. It may be heresy, but C++ excels at these sort of > low level classes: operator overloading allows you to use a very natural > syntax and because they can be stack based and C++ supports inlining > they're as efficient as raw structs. Is it possible that some simple C++ > classes will be added to the AppKit? Or are there a lot of people not using > Objective-C++? I'm curious, what "operators" would you personaly choose to overload for moveTo: and sizeTo:? Somehow myRectangle*location seems a little cryptic. But [myRectangle moveTo:location] can be understood even by a non-programmer. There's also about 75 or more other methods I find usefull for manipulating rectangles. What operators would you use for those? I pity the programmer that inherits code written in the way you suggest. -- Steve Dekorte - OpenStep consultant - San Francisco
From: jesjones@halcyon.com (Jesse Jones) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: Wed, 03 Sep 1997 20:14:17 -0700 Organization: Edmark Message-ID: <jesjones-ya02408000R0309972014170001@news.halcyon.com> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <x7afhuyivf.fsf@tom.ics.ele.tue.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit In article <x7afhuyivf.fsf@tom.ics.ele.tue.nl>, tiggr@ics.ele.tue.nl (Pieter Schoenmakers) wrote: > In article <jesjones-ya02408000R0209972131290001@news.halcyon.com> jesjones@halcyon.com (Jesse Jones) writes: > > Is it possible that some simple C++ classes will be added to the > AppKit? > > NSPoint et al already are what you desire; a struct is after all nothing > but a class with all members public. --Tiggr Does NSPoint have constructors? Can you add NSPoint's? Multiply by a scalar? Does it have comparison operators? --Jesse
From: ga@ed4u.com (G. Apple) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 4 Sep 1997 03:37:39 GMT Organization: Advanced Communications Engineering, Inc. Message-ID: <5ulae3$imb$1@gte1.gte.net> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> In article <5ul48j$3dj$1@owl.slip.net>, Steve Dekorte <dekorte@slip.net> wrote: > I'm curious, what "operators" would you personaly choose > to overload for moveTo: and sizeTo:? > > Somehow myRectangle*location seems a little cryptic. > But [myRectangle moveTo:location] can be understood even by a non-programmer. > > There's also about 75 or more other methods I find usefull for > manipulating rectangles. What operators would you use for those? > > I pity the programmer that inherits code written in the way you suggest. Then you should look at ACS which I previously mentioned. The overridden arithmetic operators "+", "-", "+=", and "-=" allow you to add a point to a rectangle, translating it by the specified point vector. It's a heck of a lot more intuitive than "[myRectangle moveTo:location]". -- G. Gordon Apple, PhD The Ed4U Project Advanced Communications Engineering, Inc. Redondo Beach, CA ga@ed4u.com
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 4 Sep 1997 02:03:11 GMT Organization: Apple Computer, Inc. Message-ID: <5ul4sv$llm$1@news.apple.com> References: <5uklcd$dmp$1@concorde.ctp.com> Thomas Engel writes > Ok..why I ran into this is simply because: > > [myDestField setStringValue:[mySourceField stringValue]] > > Bang'ed on me. And IMHO the value which is returned by an objects > "getter" should be a valid parameter for the corresponding "setter". > Besides that this behavior is not documented for NSCell or NSControl. OK, now I'm really confused. Looking at the code, I can't see how stringValue would ever return nil. I mean, it's explicitly checked for in about 4 different places, and replaced with @""... ..unless the mySourceField variable in the line above is nil. *Then* you'd get a nil "return value" from stringValue. In which case, the behavior is as expected, I guess. (This is one feature of Objective-C I have a real issue with. If I had a dime for each bug I've found that indirectly resulted from messaging a nil object...) -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: 3 Sep 1997 20:58:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B0338518-199D30@206.165.44.50> References: <19970903204855473347@lc019.zianet.com> nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.misc, nntp://news.primenet.com/comp.sys.mac.system MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Your conclusions are, pardon my Anglo-Saxon, total BULLSHIT. You have bought Steve Jobs' "Big Lie." Read what Henry Knorr says at: <http://www8.zdnet.com/macweek/mw_1134/op_decade.html> "Apple relies on half-truths and omissions to justify licensing line" Relevant quote: "To buttress their argument, they noted that the licensees have been paying only about $50 per system for their Mac OS licenses. They neglected to mention the separate hardware license fees the cloners have also been paying -- in some cases, more than $180 per unit. (Note that most of this money goes straight to Apple's bottom line, since there's no cost of goods nor any new R&D expenditure required.) But the biggest flaw in Apple's case is that it is based on the original licensing contracts; somehow company spokespeople seem to have forgotten that the clone vendors already agreed to new and very different terms. These agreements were hammered out in early June, and by all accounts they provided for significant increases in license fees -- a two- to 10-fold increase, depending on the model, relative to previous combined hardware and OS royalties, according to two major licensees. " That's a new fee structure of $100 to OVER FIVE HUNDRED DOLLARS per box, if Henry is to be trusted, that the clone-makers ALREADY had agreed to, according to Gil Amelio in a public interview. Do you have anything good to say now? > Hello. > > Anything can backfire. > > I saw a post that implied that Jobs complained that 99% of > PowerComputing customers were former Apple Macintosh users. If so, > only > 1% of PowerComputing customers had conceivably *never* touched a > Macintosh before. Consequently, PowerComputing was inadvertently (I am > NOT saying intentionally) "pirating" customers from Apple. Ideally, the > clones should have attracted Wintel users to Mac platform. > > The bottom line is that Apple's initial approach to cloning was > evidently flawed. I'll blame Apple for its clumsy clone planning but > I'll also commend Apple for listening to our needs and give cloning a > try. People got hurt in this PowerComputing cloning episode but I'll > also say it could have been lawsuit city. > > If the above information is true I think Jobs acted properly in Apple's > (both the companies' and the stockholder's) interest. > ------------------------------------------------------------------------- Build a better mousetrap, yes, but then market it like crazy, or instead of catching mice you'll only collect dust. -David Yeargin -------------------------------------------------------------------------
From: nospam+next@luomat.peak.org (Timothy J. Luoma) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Re: setting the environment variable for PERL5LIB not working Date: 4 Sep 1997 03:48:19 GMT Organization: http://www.peak.org/~luomat/ Message-ID: <5ulb23$hpg$1@ha1.rdc1.nj.home.com> References: <B0336640-69F0@141.214.128.36> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: comrade@umich.edu In <B0336640-69F0@141.214.128.36> "Robert A. Decker" wrote: > > > How can I set it up so that all users have a PERL5LIB environment > variable? When I set a line for login.std inside /etc to: > set PERL5LIB=/usr/lib/perl5 > it doesn't seem to work. try setenv PERL5LIB=/usr/lib/perl5 or change their shells to a better shell, like zsh (http://www.peak.org/zsh/) and use PERL5LIB=/usr/lib/perl5; export PERL5LIB TjL -- "NeXT continually embarrassed the rest of the industry for having workable, shipping, and high-quality answers before most people knew the questions." -- Gary Longsine, full text at: http://www.stepwise.com/Articles/Editorial/TheHolyGrail.html
Subject: Re: Objective-C Fragile Base Class Newsgroups: comp.sys.next.programmer,comp.lang.objective-c References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> In-Reply-To: <5ulae3$imb$1@gte1.gte.net> From: marco@sente.ch (Marco Scheurer) Organization: Sen:te Message-ID: <340e6ca4.0@epflnews.epfl.ch> Date: 4 Sep 97 08:09:08 GMT On 09/04/97, G. Apple wrote: >In article <5ul48j$3dj$1@owl.slip.net>, Steve Dekorte <dekorte@slip.net> wrote: >> I'm curious, what "operators" would you personaly choose >> to overload for moveTo: and sizeTo:? >> >> Somehow myRectangle*location seems a little cryptic. >> But [myRectangle moveTo:location] can be understood even by a non-programmer. >> >> There's also about 75 or more other methods I find usefull for >> manipulating rectangles. What operators would you use for those? >> >> I pity the programmer that inherits code written in the way you suggest. > > Then you should look at ACS which I previously mentioned. The >overridden arithmetic operators "+", "-", "+=", and "-=" allow you to add >a point to a rectangle, translating it by the specified point vector. >It's a heck of a lot more intuitive than "[myRectangle moveTo:location]". Oh yes? I have never seen a geometry text using + like that. (aRectangle + aPoint) could just as well be understood as the "union" of aRectangle and aPoint. It is just as wrong as your interpretation. There is nothing intuitive in using "+" to translate a rectangle, it is just a meaningless C++ shortcut and abus de langage. There is nothing intuitive in using "moveTo:" to translate a rectangle, it is just readable and understable without guesswork. Intuitiveness has nothing to do here. -- Marco Scheurer (marco@sente.ch) Sen:te
From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Newsgroups: comp.sys.next.programmer Subject: Re: NSColor encoding for user defaults really missing ? Date: 4 Sep 1997 10:29:43 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5um2in$s8p$1@concorde.ctp.com> References: <5ui719$i91$1@concorde.ctp.com> <5ukqr8$8vq$1@solaris.cc.vt.edu> dlehn@crib.bevc.blacksburg.va.us (David I. Lehn) wrote: >Thomas Engel (tsengel@immd9.informatik.uni-erlangen.de) wrote: >> I just spend a couple of hours working on a MiscKit extension to easiely >> store NSColors in the user defaults... >... >> ... however...the more I coded the less I could believe that NeXT did not >> already provide a way to encode and decode colors. >... >> Could somebody slap me against the head by giving me a pointer to a NeXT >> supplied function...or make me feel good by confirming that this functionaliy >> is really missing ? > >In Draw.app source ( propertyList.{hm} ): >id propertyListFromNSColor(NSColor *color); >NSColor *colorFromPropertyList(id plist, NSZone *zone); > >MiscKit could use a subclass of NSCoder to deal with property lists. >Something like what GNUstep was heading for. > Thanx for all the replies which confirmed my suspicion that this is really missing (and propably will not be in DR1 too) Don already has my NSColor (NSUserDefaults) category. What be basically do is to use a string encoding that is pretty much identical to [myColor description]; All you do is [myDefaults setColor:... forKey:...]; [myDefaults colorForKey:...]; and we have the main NSColor additions: + (NSColor *)colorWithStringRepresentation:(NSString *)aString; - (NSString *)stringRepresentation; and it works for all standard Openstep color spaces. The next MiscKit should be out in one or two weeks according to Don. Aloha Tomi
From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 4 Sep 1997 10:39:58 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5um35u$sg5$1@concorde.ctp.com> References: <5uklcd$dmp$1@concorde.ctp.com> <5ul4sv$llm$1@news.apple.com> MaRK_BeSSeY@Apple.Com (Mark Bessey) wrote: >Thomas Engel writes >> Ok..why I ran into this is simply because: >> >> [myDestField setStringValue:[mySourceField stringValue]] >> >> Bang'ed on me. And IMHO the value which is returned by an objects >> "getter" should be a valid parameter for the corresponding "setter". >> Besides that this behavior is not documented for NSCell or NSControl. > >OK, now I'm really confused. Looking at the code, I can't see how >stringValue would ever return nil. I mean, it's explicitly checked for in >about 4 different places, and replaced with @""... > Uuups. Ok. The problem was a couple of month ago and I did not retest the bahavior prior to my post. Sorry. I propably should go back and recheck the stuff I was doing...maybe it was a NSForm and not TextField. >..unless the mySourceField variable in the line above is nil. *Then* >you'd get a nil "return value" from stringValue. In which case, the >behavior is as expected, I guess. This was not the problem sicne it was the first thing I checked. If I have the time I'll try to reproduct that one. But my main point was the behavior of NSColorWell. aloha Tomi P.S: >(This is one feature of Objective-C I have a real issue with. If I had a >dime for each bug I've found that indirectly resulted from messaging a nil >object...) > I consider the *nil* messaging a feature and agree with NeXTs ObjC documentation which also mentiones that messaging nil objects is perfectly valid and useful.
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 04 Sep 1997 12:36:32 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7en75cqmn.fsf@tom.ics.ele.tue.nl> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <x7afhuyivf.fsf@tom.ics.ele.tue.nl> <jesjones-ya02408000R0309972014170001@news.halcyon.com> In-reply-to: jesjones@halcyon.com's message of Wed, 03 Sep 1997 20:14:17 -0700 In article <jesjones-ya02408000R0309972014170001@news.halcyon.com> jesjones@halcyon.com (Jesse Jones) writes: Does NSPoint have constructors? Can you add NSPoint's? Multiply by a scalar? Does it have comparison operators? Are you not satisfied with the NSPoint class that I provide you with? Sorry, pal, this is C++ and there is nothing you can do about deficiencies of a class over which you have no control. --Tiggr
Date: Thu, 04 Sep 1997 10:44:56 -0400 From: joe.ragosta@dol.net (Joe Ragosta) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Message-ID: <joe.ragosta-ya02408000R0409971044560001@news.dol.net> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> Organization: Graver Chemical Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit In article <B0338518-199D30@206.165.44.50>, "Lawson English" <english@primenet.com> wrote: > That's a new fee structure of $100 to OVER FIVE HUNDRED DOLLARS per box, if > Henry is to be trusted, that the clone-makers ALREADY had agreed to, > according to Gil Amelio in a public interview. Your own statement contains the answer. "IF Henry is to be trusted". Typically, his editorials have been full enough of FUD to qualify them for the National Enquirer. -- Regards, Joe Ragosta See the Complete Macintosh Advocacy Site http://www.dol.net/~Ragosta/complmac.htm
From: <sschaper@inlink.com> Subject: Re: "The OpenStep Specification" References: <340A6D61.6A88@rs6000.cmp.ilstu.edu> Message-ID: <01bcb940$8330f300$1978c4ce@mercury.inlink.com> Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.mac.advocacy Date: Thu, 04 Sep 1997 09:48:22 -0500 With Steve Jobs revealing that he has not yet escaped the Reality Distortion Field, Rhapsody may be irrelevant. GNUStep/Hurd may be the OS of the future of those who do not wish to be Microserfs.
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 04 Sep 1997 17:25:49 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7d8mpcd8i.fsf@tom.ics.ele.tue.nl> References: <5uklcd$dmp$1@concorde.ctp.com> <5ul4sv$llm$1@news.apple.com> <5um35u$sg5$1@concorde.ctp.com> In-reply-to: tsengel@immd9.informatik.uni-erlangen.de's message of 4 Sep 1997 10:39:58 GMT In article <5um35u$sg5$1@concorde.ctp.com> tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) writes: I consider the *nil* messaging a feature and agree with NeXTs ObjC documentation which also mentiones that messaging nil objects is perfectly valid and useful. Messaging nil is more often an error than a feature, to the extent that in the code I write, messaging nil is forbidden. Messaging nil returns bogus values if the size of the return type does not fit an int. Messaging nil does not do checks on the kind of messages you allow to be sent to nil. Messaging nil either is an error, or aids you in finding bogus nil messaging, in which case `normal nil messaging' will get in the way and thus should be forbidden as well. Messaging nil is a bad idea. --Tiggr
From: ga@ed4u.com (G. Apple) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 4 Sep 1997 16:35:06 GMT Organization: Advanced Communications Engineering, Inc. Message-ID: <5umnvq$ju9$1@gte2.gte.net> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> In article <340e6ca4.0@epflnews.epfl.ch>, marco@sente.ch (Marco Scheurer) wrote: > Oh yes? I have never seen a geometry text using + like that. > > (aRectangle + aPoint) could just as well be understood as the "union" > of aRectangle and aPoint. It is just as wrong as your interpretation. > > There is nothing intuitive in using "+" to translate a rectangle, it > is just a meaningless C++ shortcut and abus de langage. Considering that a rectangle is usually defined in terms of two points, one of which is usually defined as the location, addition of these vectors to show translation is much more intuitive to a "non-programmer" than the computereez that you suggested. It's obvious that you've never spent time working with vectors, vector spaces, or abstract algebra. The concept of an "operator" is something that is well understood by people versed in mathematics, which often unfortunately does not include programmers. BTW, this has been used by MacApp for many years. Release 13 splits-off these classes into the universally-usable Apple Class Suites and adds many more classes that handle other things such as multi-threading, networking, autopointers, reference counting, and a lot of other good stuff. It uses the power of C++ to its maximum advantage -- including appropriate use of operator overloading and multiple inheritance. -- G. Gordon Apple, PhD The Ed4U Project Advanced Communications Engineering, Inc. Redondo Beach, CA ga@ed4u.com
From: "Robert A. Decker" <comrade@umich.edu> Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Re: setting the environment variable for PERL5LIB not working Date: 4 Sep 97 12:36:16 -0400 Organization: University of Michigan ITD News Server Message-ID: <B0345BCE-2D1A1D@141.214.128.36> References: <199709040348.XAA07076@kira.peak.org> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: "Timothy J. Luoma" <nospam+next@luomat.peak.org> nntp://news.itd.umich.edu/comp.sys.next.programmer On Wed, Sep 3, 1997 11:48 PM, Timothy J. Luoma <mailto:nospam+next@luomat.peak.org> wrote: >setenv PERL5LIB=/usr/lib/perl5 > THanks, that fixed it. However, a new problem has come up. I seem to be missing vars.pm and about half of the other lib files i thought were standard. Should I download perl and install? I'm using the standard installation that came with the system. thanks, rob -- <mailto: "Robert A. Decker" comrade@umich.edu> <http://hmrl.cancer.med.umich.edu/Rob/index.ssi> Programmer Analyst - Health Media Research Lab University of Michigan Comprehensive Cancer Center "Get A Life" quote #3: "I'm not interested in acting. I just have the urge to perform onstage in an animal costume. There's a difference you know." -Chris Elliott
From: Matthew_Seaman@plsys.co.uk (Matthew Seaman) Newsgroups: comp.sys.next.sysadmin,comp.sys.next.programmer Subject: Re: setting the environment variable for PERL5LIB not working Date: 4 Sep 1997 17:23:18 GMT Organization: P&L Systems Message-ID: <5umqq6$j5f$3@ironhorse.plsys.co.uk> References: <199709040348.XAA07076@kira.peak.org> <B0345BCE-2D1A1D@141.214.128.36> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: comrade@umich.edu In <B0345BCE-2D1A1D@141.214.128.36> "Robert A. Decker" wrote: > THanks, that fixed it. However, a new problem has come up. I seem to be > missing vars.pm and about half of the other lib files i thought were > standard. Should I download perl and install? I'm using the standard > installation that came with the system. Definitely. The current version is perl 5.00403, many bug-fixes and enhancements ahead of the 5.001 version shipped with OpenStep. For instance, the IO, CGI and CPAN modules are now parts of the default distribution. This version of Perl compiles and installs pretty smoothly, except I found that I had to edit the cflags script to make sure that toke.c compiled without optimization --- the compiler crashes on that file otherwise: five-bells:...local/src/perl5.004_03:% diff -u cflags.orig cflags --- cflags.orig Thu Sep 4 18:21:34 1997 +++ cflags Thu Sep 4 18:20:37 1997 @@ -84,7 +84,7 @@ scope) ;; sv) ;; taint) ;; - toke) ;; + toke) optimize="-O0" ;; usersub) ;; util) ;; *) ;; Matthew [Posted and Mailed] -- Certe, Toto, sentio nos in Kansate nin iam adesse. Matthew Seaman P&L Systems, 12 The Broadway, Amersham, Bucks., HP7 0HP, UK Tel: +44 1494 432422 Fax: +44 1494 432478
From: Chris Roehrig <croehrig@House.ORG> Newsgroups: comp.sys.next.programmer Subject: I'm not understanding something about user vs. kernel memory spaces Date: 4 Sep 1997 18:55:12 GMT Organization: Computer Science, University of B.C., Vancouver, B.C., Canada Message-ID: <5un06g$1eo$1@nnrp.cs.ubc.ca> I'm working on a Matrox Meteor driver, and I'm somewhat confused about how the driver deals with user-space addresses. From what I've managed to deduce, the UNIX entry points run in the context of the user task. So they are using the virtual memory map of the user's process, right? However, they also refer to variables and memory (locks, etc) that are local to the driver and therefore in the kernel's IOTask memory space. How the heck does this work? e.g. how is this possible...? my_ioctl( dev_t dev, int cmd, void *data, int flag ) { ... bcopy( (void*)data, (void *)&my_struct, sizeof(my_struct)); where my_struct is local to the driver, is accessed by other driver routines like commandRequestOccurred, and presumable exists in the kernel's memory space. I don't really understand how the kernel knows what addresses are in user space and what are in the kernel's space. If you can do the above bcopy, then what are copyin and copyout for? Thanks, Chris -- Chris Roehrig croehrig@House.ORG Neuroscience and Computer Science at University of British Columbia, Vancouver http://www.House.ORG/chris http://www.sns.cs.ubc.ca/chris
From: "John Huang" <john.nospam@scenarist.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: 4 Sep 1997 19:12:01 GMT Organization: North Bay Network, Inc. news server - not responsible for content Message-ID: <01bcb967$7a41bd70$b55db8ce@paper> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> > > But the biggest flaw in Apple's case is that it is based on the original > licensing contracts; somehow company spokespeople seem to have forgotten > that the clone vendors already agreed to new and very different terms. > These agreements were hammered out in early June, and by all accounts they > provided for significant increases in license fees -- a two- to 10-fold > increase, depending on the model, relative to previous combined hardware > and OS royalties, according to two major licensees. " > > That's a new fee structure of $100 to OVER FIVE HUNDRED DOLLARS per box, if > Henry is to be trusted, that the clone-makers ALREADY had agreed to, > according to Gil Amelio in a public interview. > First, nobody knows what the fee structure really is. If i remembered correctly, 2 to 10 times (that means 100 to 500 dollars) per license was What Apple thought was fair, and demanded, but the agreement was much lower.....50 to 200 dollars per license, i remember reading somewhere. Also, some articles (including MacWeeks) mentioned that Power demanded 20 dollars per license because that was what PC makers pay MS. Anyway, MacWeek gets their information from rumor sources, not much different from MacOS Rumors site. Didn't Dan Crabb have an article in which he swears that Amelio resigned for health reasons right after Amelio got boosted? John
From: don@globalobjects.com (Don Yacktman) Newsgroups: comp.sys.next.programmer Subject: Re: NSColor encoding for user defaults really missing ? Date: 4 Sep 1997 19:12:46 GMT Organization: Global Objects Inc. Message-ID: <5un17e$q57$1@news.xmission.com> References: <5ui719$i91$1@concorde.ctp.com> <5ukqr8$8vq$1@solaris.cc.vt.edu> <5um2in$s8p$1@concorde.ctp.com> tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) wrote: > dlehn@crib.bevc.blacksburg.va.us (David I. Lehn) wrote: > >MiscKit could use a subclass of NSCoder to deal with property lists. > >Something like what GNUstep was heading for. > > Don already has my NSColor (NSUserDefaults) category. Yup. It will definitely be in the next release; I've already added it to my local MiscKit code repository. > The next MiscKit should be out in one or two weeks according to Don. I hope to have the release out by Saturday. If I slip, then it will be at least two more weeks before it goes out. Let's hope for the best... -- Later, -Don Yacktman don@misckit.com <a href="http://www.misckit.com/don.html">My home page</a>
From: Chris Roehrig <croehrig@House.ORG> Newsgroups: comp.sys.next.programmer Subject: How to do DMA directly to user memory? Date: 4 Sep 1997 19:16:29 GMT Organization: Computer Science, University of B.C., Vancouver, B.C., Canada Message-ID: <5un1ed$1o3$1@nnrp.cs.ubc.ca> I'm trying to set up DMA directly to user memory (about 100K's worth) from a UNIX-style entry point (ioctl), and I don't know how to wire-down the user memory. This is the best I can figure out... // get physical address of user buffer... // But this only works if the memory is already wired down! IOPhysicalFromVirtual( IOVmTaskCurrent(), userBufAddr, &physAddr ); // map it into the driver's memory space... IOMapPhysicalIntoIOTask( physAddr, userBufSize, &myVirtAddr ); // ... and wire it down. // this only works from kernel servers? What is the first parm? kern_serv_wire_range( ..., myVirtAddr, userBufSize ); // send a Mach msg to IOThread to start DMA to physAddr... ... Thanks, Chris -- Chris Roehrig croehrig@House.ORG Neuroscience and Computer Science at University of British Columbia, Vancouver http://www.House.ORG/chris http://www.sns.cs.ubc.ca/chris
From: scottm@nic.com (Scott) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: Thu, 04 Sep 1997 15:58:05 -0400 Organization: PETA (People for the Eating of Tasty Animals) Message-ID: <scottm-ya02408000R0409971558050001@news.erols.com> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> <joe.ragosta-ya02408000R0409971044560001@news.dol.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit In article <joe.ragosta-ya02408000R0409971044560001@news.dol.net>, joe.ragosta@dol.net (Joe Ragosta) wrote: >In article <B0338518-199D30@206.165.44.50>, "Lawson English" ><english@primenet.com> wrote: > >> That's a new fee structure of $100 to OVER FIVE HUNDRED DOLLARS per box, if >> Henry is to be trusted, that the clone-makers ALREADY had agreed to, >> according to Gil Amelio in a public interview. > >Your own statement contains the answer. "IF Henry is to be trusted". >Typically, his editorials have been full enough of FUD to qualify them for >the National Enquirer. > Hmmm, let's see. I can either trust Henry or Jobs. Hmmm, guess who I choose. Hint: It's not Jobs. -- -------------------------------------------------------------------------- Scott Maxwell - scottm (at) nic (dot) com Finger me for my public pgp key or http://www.nic.com/~scottm/key.html Remember when SPAM was the world's favorite mystery meat? Surfing is fast. The Internet is not. Therefore I "Sloth The Net"
Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> <5umnvq$ju9$1@gte2.gte.net> From: noone@iphysiol.unil.ch (Sean Hill) Message-ID: <340f1650.0@cisun2000.unil.ch> Date: 4 Sep 97 20:13:04 GMT ga@ed4u.com (G. Apple) wrote: > In article <340e6ca4.0@epflnews.epfl.ch>, marco@sente.ch (Marco Scheurer) wrote: > > Oh yes? I have never seen a geometry text using + like that. > > > It's obvious that you've never spent time > working with vectors, vector spaces, or abstract algebra. The concept of > an "operator" is something that is well understood by people versed in > mathematics, which often unfortunately does not include programmers. > That's funny!!! If you knew Marco, you would not say that... Ciao! -sean
From: jsturm@zenacomp.zenacomp.com (Jeff Sturm) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Followup-To: comp.sys.next.programmer,comp.lang.objective-c Date: 4 Sep 1997 18:32:21 -0400 Organization: ZenaComp Incorporated Message-ID: <5unctl$e0n@zenacomp.zenacomp.com> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> <5umnvq$ju9$1@gte2.gte.net> G. Apple (ga@ed4u.com) wrote: : In article <340e6ca4.0@epflnews.epfl.ch>, marco@sente.ch (Marco Scheurer) wrote: : > Oh yes? I have never seen a geometry text using + like that. : > : > (aRectangle + aPoint) could just as well be understood as the "union" : > of aRectangle and aPoint. It is just as wrong as your interpretation. : > : > There is nothing intuitive in using "+" to translate a rectangle, it : > is just a meaningless C++ shortcut and abus de langage. : Considering that a rectangle is usually defined in terms of two points, : one of which is usually defined as the location, addition of these vectors : to show translation is much more intuitive to a "non-programmer" than the : computereez that you suggested. It's obvious that you've never spent time : working with vectors, vector spaces, or abstract algebra. The concept of : an "operator" is something that is well understood by people versed in : mathematics, which often unfortunately does not include programmers. Whoa. I have a BS in mathematics, and yet I program computers every day. I am aware that there is a rigorous notation for mathematical operators, so mathematicians can understand one another. I'm also aware that it is dangerous at best to try associating math operators with programming languages. Part of the problem is that many of my favorite math operators just can't be typed on my lowly ASCII keyboard. Besides, in the entire history of programming languages, no preferential set of binary and unary operators has ever emerged, likely because there isn't one--it varys by personal taste. That's why the two of you can both quite rightly believe in what you say and yet may never agree. That said, I have witnessed too many variations and abuses of simple operators. You must realize that 'moveTo:' is an operator too, it doesn't have to be short or made of punctuation symbols (unless your programming language requires it) to be on equal mathematical footing. Personally, I like moveTo: because it is a whole lot more descriptive, and I don't have to think about what '+' means in that context. -- Jeff Sturm ZenaComp Incorporated jsturm@zenacomp.com
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 05 Sep 1997 00:43:41 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7wwkwem3m.fsf@tom.ics.ele.tue.nl> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> <5umnvq$ju9$1@gte2.gte.net> In-reply-to: ga@ed4u.com's message of 4 Sep 1997 16:35:06 GMT In article <5umnvq$ju9$1@gte2.gte.net> ga@ed4u.com (G. Apple) writes: Considering that a rectangle is usually defined in terms of two points, one of which is usually defined as the location, [...] Considering that in the AppKit, a rectangle is defined by its bottom-left corner (an NSPoint) and its width and height (an NSSize), your remarks do not apply. --Tiggr
From: "Tim Priest" <tpriest@metz.une.edu.au> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: 5 Sep 97 09:54:36 +0000 Organization: University of New England, NSW, Australia Message-ID: <B0358762-78907@129.180.65.35> References: <joe.ragosta-ya02408000R0409971044560001@news.dol.net> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit nntp://News.une.edu.au/comp.sys.next.advocacy, nntp://News.une.edu.au/comp.sys.mac.programmer.misc, nntp://News.une.edu.au/comp.sys.next.programmer, nntp://News.une.edu.au/comp.sys.mac.misc, nntp://News.une.edu.au/comp.sys.mac.system On Thu, 4 Sep 1997 2:44 PM, Joe Ragosta <mailto:joe.ragosta@dol.net> wrote: >> Henry is to be trusted, that the clone-makers ALREADY had agreed to, >> according to Gil Amelio in a public interview. > >Your own statement contains the answer. "IF Henry is to be trusted". >Typically, his editorials have been full enough of FUD to qualify them for >the National Enquirer. > Well I rwad the interview with on Macaddict Gil and he states that a tentative agreement was worked out two weeks before Jobs took over. So it has nothing to do with Henry. Just Steve! I am wondering! Is there really a Joe Ragosta? Or is it a clever aliasw for STEVE JOBS! Ha, yes, thats it. Nothing else explains the blind faith and evangelism shown by this so-called Joe Ragosta! PS: Only joking Joe! I wish I could have your faith in Apple. Tim Priest --------------------------------------------------- This message was created and sent using the Cyberdog Mail System ---------------------------------------------------
From: rmcassid@uci.edu (Bob Cassidy) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: Thu, 04 Sep 1997 17:17:33 -0700 Organization: UC Irvine Message-ID: <rmcassid-0409971717330001@dante.eng.uci.edu> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> <joe.ragosta-ya02408000R0409971044560001@news.dol.net> In article <joe.ragosta-ya02408000R0409971044560001@news.dol.net>, joe.ragosta@dol.net (Joe Ragosta) wrote: > In article <B0338518-199D30@206.165.44.50>, "Lawson English" > <english@primenet.com> wrote: > > > That's a new fee structure of $100 to OVER FIVE HUNDRED DOLLARS per box, if > > Henry is to be trusted, that the clone-makers ALREADY had agreed to, > > according to Gil Amelio in a public interview. > > Your own statement contains the answer. "IF Henry is to be trusted". > Typically, his editorials have been full enough of FUD to qualify them for > the National Enquirer. $500 huh. A $5000 high-end box should net Apple $1200 assuming a 20% margin. Power would need to bring in at least two new customers to justify the license. -Bob Cassidy
From: Steve Dekorte <dekorte@slip.net> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: 5 Sep 1997 00:25:43 GMT Organization: Slip.Net (http://www.slip.net) Message-ID: <5unji7$nf0$1@owl.slip.net> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> In comp.lang.objective-c G. Apple <ga@ed4u.com> wrote: > Steve Dekorte <dekorte@slip.net> wrote: > > There's also about 75 or more other methods I find usefull for > > manipulating rectangles. What operators would you use for those? > > > > I pity the programmer that inherits code written in the way you suggest. > Then you should look at ACS which I previously mentioned. The > overridden arithmetic operators "+", "-", "+=", and "-=" allow you to add > a point to a rectangle, translating it by the specified point vector. > It's a heck of a lot more intuitive than "[myRectangle moveTo:location]". Ok, you've covered the moveBy: method. Now you only have 74 more methods to make operators for. I'm very interested in seeing your choices. Even if just for moveTo:, sizeTo:, scaleBy:, rotateBy:, fitInside:, and instercetionWith:. -- Steve Dekorte - OpenStep consultant - San Francisco
Newsgroups: comp.sys.next.programmer Subject: pid of NSTask? Message-ID: <3dgOoYuhHGqY@cc.usu.edu> From: edx@cc.usu.edu Date: 4 Sep 97 18:22:06 MDT Distribution: world MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Is there a way to get the pid of a program which was started using the launch method of NSTask? I don't see a "pid" method in the NSTask method list. Thanks for any pointers. edx@cc.usu.edu
From: jesjones@halcyon.com (Jesse Jones) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: Thu, 04 Sep 1997 20:44:19 -0700 Organization: Edmark Message-ID: <jesjones-ya02408000R0409972044190001@news.halcyon.com> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit In article <5ul48j$3dj$1@owl.slip.net>, Steve Dekorte <dekorte@slip.net> wrote: > In comp.lang.objective-c Jesse Jones <jesjones@halcyon.com> wrote: > > In article <872781306.21356@dejanews.com>, David Stes <stes@can.nl> wrote: > > > > The Point and Rectangle classes of ICpak101 are nice. > > > > > > In ICpak201 (for X/Windows) you have then nifty methods > > > like -moveTo:aPoint or -sizeTo:aPoint. > > > But effiency is important, especially for something as commonly used as > > points and rectangles. It may be heresy, but C++ excels at these sort of > > low level classes: operator overloading allows you to use a very natural > > syntax and because they can be stack based and C++ supports inlining > > they're as efficient as raw structs. Is it possible that some simple C++ > > classes will be added to the AppKit? Or are there a lot of people not using > > Objective-C++? > > I'm curious, what "operators" would you personaly choose > to overload for moveTo: and sizeTo:? > > Somehow myRectangle*location seems a little cryptic. > But [myRectangle moveTo:location] can be understood even by a non-programmer. You're quite right: myRectangle*location is awfully cryptic. That's why good C++ programmers are very careful about how they use operator overloading: it's too easy to obfuscate the meaning of the code. OTOH there are a few objects (mostly from mathematics) where operator overloading is a boon to clarity. My point class is modeled after MacApp's. It allows you to add and subtract points, multiply and divide points by a scalar, and compare points. I would claim that these are completely natural and intuitive operations. My rectangle class OTOH is perhaps not quite so intuitive. It overloads operator[] (with the subscript being an enumeration with values topLeft and botRight), operator+ and - (for translattion using a point), operator & and | (for intersection and union with another rectangle), and operator== and !=. Here's some code I recently wrote using these operators: XPoint loc = bounds[topLeft] + mLocation + XPoint(kCellIndent, kCellIndent) - revealed[topLeft]; Is this not much clearer than the non-overloaded version? > There's also about 75 or more other methods I find usefull for > manipulating rectangles. What operators would you use for those? The only place I have ever used operator overloading is for mathematical objects (points, rectangles, vectors, arrays, complex numbers, etc). I've heard about people overloading operator+= to add a pane to a view or a new element to a list, but I'd consider that an abuse of overloading. If I had 75 methods in my rect class you can be damn sure that very few of them would be operators. > I pity the programmer that inherits code written in the way you suggest. I pity the person who can't see that it's possible to use a dangerous tool wisely. --Jesse
From: jesjones@halcyon.com (Jesse Jones) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class Date: Thu, 04 Sep 1997 20:51:43 -0700 Organization: Edmark Message-ID: <jesjones-ya02408000R0409972051430001@news.halcyon.com> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <x7afhuyivf.fsf@tom.ics.ele.tue.nl> <jesjones-ya02408000R0309972014170001@news.halcyon.com> <x7en75cqmn.fsf@tom.ics.ele.tue.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit In article <x7en75cqmn.fsf@tom.ics.ele.tue.nl>, tiggr@ics.ele.tue.nl (Pieter Schoenmakers) wrote: > In article <jesjones-ya02408000R0309972014170001@news.halcyon.com> jesjones@halcyon.com (Jesse Jones) writes: > > Does NSPoint have constructors? Can you add NSPoint's? Multiply by a > scalar? Does it have comparison operators? > > Are you not satisfied with the NSPoint class that I provide you with? > Sorry, pal, this is C++ and there is nothing you can do about deficiencies > of a class over which you have no control. --Tiggr What class? It's a plain struct with no methods, right? If so, then no I'm not satisfied. C++ does a great job at making simple objects like points and rects a pleasure to work with. I don't see why the AppKit can't take limited advantage of C++ (I heard at WWDC (or maybe MacHack) that Apple's Objective-C is going to use the C++ exception model so it seems like the AppKit can count on being compiled with C++). And, as it turns out, you're wrong. Given an NSPoint that's a simple C-style struct I can overload the global operators to work with NSPoint (although I can't implement += and -= properly). --Jesse
From: tom@sumoweb.com (Thomas McCarthy) Newsgroups: comp.sys.next.programmer Subject: Can't make msql -- why? Date: 5 Sep 1997 04:11:54 GMT Organization: Pacific Bell Internet Services Message-ID: <5uo0qa$6im$1@nnrp1.snfc21.pbi.net> Hi, This is probably an absurdly basic question, but I am only trying to use make (which I know nothing about) because I haven't come across a precompiled version of msql, which I would like to run on my slab. I followed the instructions that came with the distribution, but after building the target directory and cd'ing into it, I get this when I try to make all: bash-2.00$ cd /Users/tom/Downloads/msql.1.0.6/targets/NeXT-4.2-Motorola bash-2.00$ make all Make: Don't know how to make Makefile.tmpl. Stop. bash-2.00$ Is there some obvious step or common problem I am unaware of? (And is there a compiled msql hiding somewhere?) Any help appreciated! -- Tom ---------- Thomas McCarthy tom@sumoweb.com
From: Ingo Feulner <ifeulner@xenon.cube.de> Newsgroups: comp.sys.next.programmer Subject: Re: NSColor encoding for user defaults really missing ? Date: 4 Sep 1997 09:31:34 GMT Organization: Private Site, =?ISO-8859-1?Q?B=F6blingen?=, Germany Message-ID: <5ulv5m$8qg$1@xenon.cube.de> References: <5ui719$i91$1@concorde.ctp.com> Mime-Version: 1.0 (FinalNews for NeXTstep; Version 0.32 / Jul 12, 1997) Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Xcanpos: shelf.01/199709081001!0021321204 Thomas Engel wrote in <5ui719$i91$1@concorde.ctp.com> on 1997-09-03 01:21:13 +0200: > Hi, > > I just spend a couple of hours working on a MiscKit extension to easiely > store NSColors in the user defaults... Here it is: -------NSUserDefaults_Color.h--------------------------------------------- #import <Foundation/NSUserDefaults.h> #import <AppKit/NSColor.h> @interface NSUserDefaults(Color) - (NSColor *)colorForKey:(NSString *)key; - (void)setColor:(NSColor *)color forKey:(NSString *)defaultName; @end -------NSUserDefaults_Color.m--------------------------------------------- #import "NSUserDefaults_Color.h" @implementation NSUserDefaults(Color) - (NSColor *)colorForKey:(NSString *)defaultName { NSColor * color; id obj; color = [NSColor blackColor]; obj = [self stringForKey:defaultName]; if (obj) { NSScanner *scanner; float red, green, blue, alpha; scanner = [NSScanner scannerWithString:obj]; [scanner scanFloat:&red]; [scanner scanFloat:&green]; [scanner scanFloat:&blue]; [scanner scanFloat:&alpha]; color = [NSColor colorWithCalibratedRed:red green:green blue:blue alpha:alpha]; } return color; } - (void)setColor:(NSColor *)color forKey:(NSString *)defaultName; { NSString * string; float red, green, blue, alpha; [[color colorUsingColorSpaceName:NSCalibratedRGBColorSpace] getRed:&red green:&green blue:&blue alpha:&alpha]; string = [NSString stringWithFormat:@"%f %f %f %f", red, green, blue, alpha]; [self setObject:string forKey:defaultName]; } @end Bye, Ingo. -- Smail: Ingo Feulner, Wolfacher Weg 19, 71034 Böblingen, Germany Email: ifeulner@xenon.cube.de "If you want to convince yourself how nice OpenStep is, try Java" - Wiliam Shipley.
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: 5 Sep 1997 01:18:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B0351358-2393B@206.165.44.43> References: <rmcassid-0409971717330001@dante.eng.uci.edu> To: "Bob Cassidy" <rmcassid@uci.edu> nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.misc, nntp://news.primenet.com/comp.sys.mac.system MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Bob Cassidy <rmcassid@uci.edu> said: > > $500 huh. A $5000 high-end box should net Apple $1200 assuming a 20% > margin. Power would need to bring in at least two new customers to > justify > the license. > > -Bob Cassidy Frankly, I'm disappointed in you, Robert. The 20% margins that are tossed around are "gross margins." There's a LOT of overhead that isn't included in that that knocks the net profit down. ON the other hand, for already-designed hardware and already created OS software, the only overhead for licensing fees has to do with the handling of the licensing and the certifying of the hardware boxes. For all practical purposes, that $500 is darned near pure profit for Apple. Not so with that $1200. You have an edu address. Go ask the accounting profs if you have doubts. If they tell you different, come back and I'll apologize. ------------------------------------------------------------------------- Build a better mousetrap, yes, but then market it like crazy, or instead of catching mice you'll only collect dust. -David Yeargin -------------------------------------------------------------------------
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 04:31:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B03540AE-CE02A@206.165.44.43> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.system, nntp://news.primenet.com/comp.sys.mac.misc MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Matthew Vaughan <see-below@not-my-address.com> said: > I think he meant that Ellison had originally attempted to buy Apple, and > convert it into an NC manufacturer. Now he gets appointed to the board > with essentially no money invested, and accomplishes the same thing. What is really fun is that I'm starting to get details about what my "Very Bad Thing" may turn out to be: JObs was told by Motorola to stop playing games with clone-makers or face losing PowerPC's from Motorola. Jobs replied that he didn't care since Net Computers could just as easily run on x86 CPUs. (dare I point out that Rhapsody is cross-platform?) Guess what? No more PowerPCs from Motorola starting in 1999. Word is that the decision is moving through channels, even as we speak. Since Rhapsody is cross-platform, who cares if Apple doesn't make PowerMacs anymore? Not Steve or Larry. They're after The NeXT BIg Thing, which is Net Computers, with high-end servers. It all becomes clear now... ++++++++++++++++++++++++++++++ HOW DO I SUBSCRIBE TO GX-TALK? --------------------------------- To subscribe to GX-TALK proceed as follows: Send E-mail to: <gx-talk-request@aimed.org> with anything in the subject line and the following command as the first (and only) line of the message body: SUBSCRIBE +++++++++++++++++++++++++++++++++
Subject: Re: Objective-C Fragile Base Class (long) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> <5umnvq$ju9$1@gte2.gte.net> In-Reply-To: <5umnvq$ju9$1@gte2.gte.net> From: marco@sente.ch (Marco Scheurer) Organization: Sen:te Message-ID: <34100283.0@epflnews.epfl.ch> Date: 5 Sep 97 13:00:51 GMT On 09/04/97, G. Apple wrote: >In article <340e6ca4.0@epflnews.epfl.ch>, marco@sente.ch (Marco Scheurer) wrote: >> Oh yes? I have never seen a geometry text using + like that. >> >> (aRectangle + aPoint) could just as well be understood as the "union" >> of aRectangle and aPoint. It is just as wrong as your interpretation. >> >> There is nothing intuitive in using "+" to translate a rectangle, it >> is just a meaningless C++ shortcut and abus de langage. > > Considering that a rectangle is usually defined in terms of two points, >one of which is usually defined as the location, addition of these vectors >to show translation is much more intuitive to a "non-programmer" than the >computereez that you suggested. It's obvious that you've never spent time >working with vectors, vector spaces, or abstract algebra. The concept of >an "operator" is something that is well understood by people versed in >mathematics, which often unfortunately does not include programmers. No kidding? So "+" has to be THE operator to translate a rectangle and moveTo: (or translateBy:) would be less acceptable? You are confusing operator and symbol. Anyway, you're wrong on both counts: mathematics and programming. MATHEMATICS In my practice of math, "+" is always used for a function with the properties of addition in Z (the set of integers), and defines a group with the set on which it operates. Generally (G, .) denotes the group of the set G with the operation "." . Practically, . is replaced with + when it is commutative and shows other analogies with the addition of real numbers: for instance the addition of complex numbers, the addition of vectors. These analogies can also, in the case of rings, refer to the properties of addition with respect to multiplication, for instance the behavior of their respective identity elements. Nevertheless, even without refering to these analogies, and not event the commutativity of addition (is your overloaded + operator commutative? I didn't think so), + should be the symbol of an operation for which closure holds (+: G x G -> G). Is it the case in the addition of a "point" (really a vector: you're also making another assumption, an implicit conversion from a point to a vector, and give a poorly defined meaning to things that have precise mathematical definitions) and a rectangle? If the answer is yes then points and rectangles belongs to the same set. The set of "geometrical elements"? The set of "things"? Try to define meaningful operations on these. So + is an acceptable symbol to designate vector addition, but not for translating rectangles. PROGRAMMING Why should a class library or a piece of C++ code be intuitive to a "non-programmer" is beyond me. The fact that your non-programmer (who's that: your user? your boss? your sponsor? you?) should be versed in mathematics also. You should realize that using (aRectangle + aPoint) to translate aRectangle is not the truth, has not much to do with math, and is just a convention used by your favorite class library. It may be obvious once you know it, but it is not intuitive, or: few people will have the same intuition as you about what a particular operator will do. Also, I wouldn't be surprised if your library uses + for the union of two rectangles. If it does, it uses + in a completely different way as the one you just described. Conventions are useful, even when not universal: used only by you, or your project team, or your company, but they are not intuitive: you have to document and learn them. Intuitive, obvious, implicit are dangerous words in software engineering, often excuses for bad practice. The use of + you described is bad practice. What do you get? It is just as much work to overload + as it is to define a new function, such as "translate", or maybe "tr" or "t", which are symbols I've found in geometry texts to denote a translation. You take risks because it is ambiguous: it could very well be understood has a form of union. You loose readability. I undertand that some people find overloading operators in general, and + in particular, at the slightest reason, cute, or cool. Unfortunately, these people often include programmers. -- Marco Scheurer (marco@sente.ch) Sen:te
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 06:12:38 -0700 Organization: Netcom Message-ID: <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> In article <B03540AE-CE02A@206.165.44.43>, "Lawson English" <english@primenet.com> wrote: > What is really fun is that I'm starting to get details about what my "Very > Bad Thing" may turn out to be: > > JObs was told by Motorola to stop playing games with clone-makers or face > losing PowerPC's from Motorola. Jobs replied that he didn't care since Net > Computers could just as easily run on x86 CPUs. (dare I point out that > Rhapsody is cross-platform?) > > Guess what? > > No more PowerPCs from Motorola starting in 1999. Word is that the decision > is moving through channels, even as we speak. Unbelievable...Steve Jobs has managed to destroy Apple in <1 year. All for his damn vision of a slotless, fanless, underpowered and overpriced machine.... > Since Rhapsody is cross-platform, who cares if Apple doesn't make PowerMacs > anymore? > > Not Steve or Larry. They're after The NeXT BIg Thing, which is Net > Computers, with high-end servers. > > It all becomes clear now... Can't somebody sue him? Apple customers? Stockholders? Motorola? Can't somebody drive him out before he kills Apple, the Mac, and Rhapsody? -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ ================================================================= ollowing command as the > first (and only) line of the message body: > > SUBSCRIBE > +++++++++++++++++++++++++++++++++ Where did you get this information?
From: d89cb@efd.lth.se (Christian Brunschen) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 13:28:12 GMT Organization: Lund Institute of Technology, Sweden Message-ID: <5up1dc$ndg$1@news.lth.se> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> NNTP-Posting-User: d89cb In article <B03540AE-CE02A@206.165.44.43>, Lawson English <english@primenet.com> wrote: >Matthew Vaughan <see-below@not-my-address.com> said: > >> I think he meant that Ellison had originally attempted to buy Apple, and >> convert it into an NC manufacturer. Now he gets appointed to the board >> with essentially no money invested, and accomplishes the same thing. > (Before I comment on the individual points raised, I must say that this article was certainly very interesting reading, and provokes much thought. Those of you who usually disregard anything Lawson says, make an exception this once at least.) > > >What is really fun is that I'm starting to get details about what my "Very >Bad Thing" may turn out to be: > >JObs was told by Motorola to stop playing games with clone-makers or face >losing PowerPC's from Motorola. Jobs replied that he didn't care since Net >Computers could just as easily run on x86 CPUs. (dare I point out that >Rhapsody is cross-platform?) > >Guess what? > >No more PowerPCs from Motorola starting in 1999. Word is that the decision >is moving through channels, even as we speak. > >Since Rhapsody is cross-platform, who cares if Apple doesn't make PowerMacs >anymore? Beginning of 1999 -- that's a little over a year away. By that time, Rhapsody Unified will have been shipping for several months, and Rhapsody Premier for almost a year. Most new application development for the Mac platform, should be happening under Rhapsody -- which means that, for new applications, indeed we do not care if we are running on PowerPC or on Intel. Those who still need PowerPC machines at that time will be those who are running MacOS applications (either 'native' in MacOS itself, or in Rhapsody's 'Blue Box'). Unless someone writes a PowerPC emulator for Intel, in which case the Blue Box might well become available there, too. Rhapsody's cross-platform abilities mean that Apple is not 'locked' into the PowerPC processor anymore, and can act freely in regard to even someone like Motorola, their chip supplier for PowerPC. 'Freely' meaning, of course, 'in Apple's self-interest', as defined by Apple's leadership. > >Not Steve or Larry. They're after The NeXT BIg Thing, which is Net >Computers, with high-end servers. Well, NC:s have been reality in different shapes for a long time. At the univerity where I study, the campus network is basically a few central file servers, and 'NC':s for people to sit at. Of course these 'NC':s are (these days) Sparc Ultra1 workstations. Yes they have local storage, but that's only for 'caching' -- all software on these machines is identical, and is even updated from the central servers. Ie, any new software that gets installed, gets installed _once_, and then 'pushed into the NC:s' cashes'. So, while they do not get their OS and software over the net _each time they need it_, they do get them from there whenever they are updated -- but through a push rather than pull action. Each user is presented with their own, customizable, interface which is identical on all these computers. Ie, user X has the setup of his choice, regardless of machine, while user Y has the setup of _her_ choice regardless of machine. Similar things have been, both possible and _done_, with NeXTSTEP / OPENSTEP machines. NetInfo in particular is rather well-tailored to this 'style' of administration -- central info that is pushed out to decentralized clone servers, from which the clients get their info. > >It all becomes clear now... Perhaps. Things remain to be seen. Also, the moniker 'Very Bad Thing' may be somewhat inaccurate. This may be 'bad' in some ways, bot _good_ in others: if Apple is forced to abandon the PowerPC platform, they will have to find another platform and support it wholeheartedly - x86. That means that people will want to be able to use vanilla x86 hardware, with its variety of cards etc, and have it all run under Rhapsody. So drivers etc for Intel should be more prolific, and should get developed faster & better, than if Apple had the PowerPC as their 'main' platform. But just as it has been said so often: Only time will tell. // Christian Brunschen
From: "Sean C." <Corliss@A1.tch.harvard.edu> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 09:26:46 -0400 Organization: Children's Hospital Message-ID: <34100896.6928@A1.tch.harvard.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Daniel L. Taylor wrote: > > In article <B03540AE-CE02A@206.165.44.43>, "Lawson English" > <english@primenet.com> wrote: > > > What is really fun is that I'm starting to get details about what my "Very > > Bad Thing" may turn out to be: > > > > JObs was told by Motorola to stop playing games with clone-makers or face > > losing PowerPC's from Motorola. Jobs replied that he didn't care since Net > > Computers could just as easily run on x86 CPUs. (dare I point out that > > Rhapsody is cross-platform?) > > > > Guess what? > > > > No more PowerPCs from Motorola starting in 1999. Word is that the decision > > is moving through channels, even as we speak. > > Unbelievable...Steve Jobs has managed to destroy Apple in <1 year. All for > his damn vision of a slotless, fanless, underpowered and overpriced > machine.... > > > Since Rhapsody is cross-platform, who cares if Apple doesn't make PowerMacs > > anymore? > > > > Not Steve or Larry. They're after The NeXT BIg Thing, which is Net > > Computers, with high-end servers. > > > > It all becomes clear now... > > Can't somebody sue him? Apple customers? Stockholders? Motorola? Can't > somebody drive him out before he kills Apple, the Mac, and Rhapsody? > > -- > Daniel L. Taylor > LTaylor7@ix.netcom.com > ================================================================= > The Heritage Foundation /// Town Hall > http://www.heritage.org/heritage/ /// http://www.townhall.com/ > ================================================================= I'd like to see a source for this rumor. I sincerely can't believe that they would do this. Sean C. Corliss@A1.tch.harvard.edu
From: Benjamin Smith <benjs@ix.netcom.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 06 Sep 1997 10:30:06 -0400 Organization: Music Mania! Message-ID: <341168EE.61A1@ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Daniel L. Taylor wrote: > > In article <B03540AE-CE02A@206.165.44.43>, "Lawson English" > <english@primenet.com> wrote: > > > What is really fun is that I'm starting to get details about what my "Very > > Bad Thing" may turn out to be: > > > > JObs was told by Motorola to stop playing games with clone-makers or face > > losing PowerPC's from Motorola. Jobs replied that he didn't care since Net > > Computers could just as easily run on x86 CPUs. (dare I point out that > > Rhapsody is cross-platform?) > > > > Guess what? > > > > No more PowerPCs from Motorola starting in 1999. Word is that the decision > > is moving through channels, even as we speak. > > Unbelievable...Steve Jobs has managed to destroy Apple in <1 year. All for > his damn vision of a slotless, fanless, underpowered and overpriced > machine.... > > > Since Rhapsody is cross-platform, who cares if Apple doesn't make PowerMacs > > anymore? > > > > Not Steve or Larry. They're after The NeXT BIg Thing, which is Net > > Computers, with high-end servers. > > > > It all becomes clear now... > > Can't somebody sue him? Apple customers? Stockholders? Motorola? Can't > somebody drive him out before he kills Apple, the Mac, and Rhapsody? > > -- Talk of using an x86 for any Apple product or MacOS machine (as opposed to Rhapsody) truly disgusts me. The PowerPC is a superior chip, is more modern and Apple should run its platform on this chip (it gives MAc superior performance in graphics) and Newtons on StrongARMs. We expect the best out of Apple and AIM, all players, must get their act together and not talk of killing this chip. The talk should be of continuing to outpace the x86 and, I guess, the vapour Merced. Not this. If Steve stands for quality and pushing the enveloppe this is truly unacceptable. Ben Smith > Daniel L. Taylor >s s s s s s s s s s s s s s s s s s s s s
From: mmalcolm crawford <malcolm@plsys.co.uk> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 14:56:26 GMT Organization: P&L Systems Message-ID: <5up6iq$ngn$1@ironhorse.plsys.co.uk> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <5up1dc$ndg$1@news.lth.se> In-Reply-To: <5up1dc$ndg$1@news.lth.se> On 09/05/97, Christian Brunschen wrote: [...] >Also, the moniker 'Very Bad Thing' may be somewhat inaccurate. This >may be 'bad' in some ways, bot _good_ in others [...] Excellent; thank you Christain. Saved me a lot of typing. Now back to doing constructive positive things to help with Apple's future... see you on DR1, RSN. Best wishes, mmalc.
From: Thomas Harte <tph1001@CL.cam.ac.uk> Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: can't COMPILE DYNAMIC LIBRARIES Date: 5 Sep 1997 15:24:02 GMT Organization: Cambridge University Message-ID: <5up86i$pob$1@lyra.csx.cam.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Maybe I am going about building shared libraries the wrong way? Feel free to put me right! I'm having difficulty compiling a dynamic library. My version of gcc does not support the -shared option so I am using ld. I get the following errors: bash-2.00$ gnumake -f makefile-shared.gnu cc++ -g -DINTEL -I/Users/tph1001/Code/include/C++libs/ezee -I/NextDeveloper/Headers/bsd/sys -c ezeeio.cc -o /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o cc++ -g -DINTEL -I/Users/tph1001/Code/include/C++libs/ezee -I/NextDeveloper/Headers/bsd/sys -c ezeeconvert.cc -o /Users/tph1001/Code/obj/C++libs/ezee/ezeeconvert.o ld -dynamic -dylib /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o /Users/tph1001/Code/obj/C++libs/ezee/ezeeconvert.o -o /Users/tph1001/Code/lib/C++libs/ezee/libezee.A.dylib.1 ld: common symbols not allowed with MH_DYLIB output format /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _RepnInfo (size 4) /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _data_sizefield (size 4) /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _dcomplexfield (size 4) /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _dsexfield (size 4) /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _dtypefield (size 4) /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _idcodefield (size 4) /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _nbandsfield (size 4) /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _ncolsfield (size 4) /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _nrowsfield (size 4) /Users/tph1001/Code/obj/C++libs/ezee/ezeeio.o definition of common _nsetsfield (size 4) gnumake: *** [/Users/tph1001/Code/lib/C++libs/ezee/libezee.A.dylib.1] Error 1 Anyone have any idea what the "ld: common symbols not allowed with MH_DYLIB output format" error message means? I have included my makefile and the area in the source where the problem lines are as indicatedd by the error message. -- thomas harte @ computer laboratory, cambridge university; tph1001@CL.cam.ac.uk; phone: +44 1223 334628; fax: 334678; http://www.CL.cam.ac.uk/users/tph1001 MIME & NeXT Mail OK. HERE's THE MAKEFILE: ==================== # ========== libraries and headers ========== MACHINE = INTEL LIB_DIR = /Users/tph1001/Code/lib/C++libs/ezee OBJ_DIR = /Users/tph1001/Code/obj/C++libs/ezee INCLUDE_DIR = /Users/tph1001/Code/include/C++libs/ezee INCLUDE = -I$(INCLUDE_DIR) -I/NextDeveloper/Headers/bsd/sys LIBS = -liostream -lm EZEELIB = $(LIB_DIR)/libezee.A.dylib CC = cc++ #.cc.o: # $(CC) $(CFLAGS) -c $< $(LIBS) CFLAGS = CFLAGS += -g -D$(MACHINE) $(INCLUDE) LD = ld LFLAGS = LFLAGS += -dynamic -dylib OBJECTS = \ ezeeio.o \ ezeeconvert.o # ========== Building rules for Makefile ========== # # ========== .so: shared object library========== $(EZEELIB): $(EZEELIB).1 rm -f $(EZEELIB); ln -s $(EZEELIB).1 $(EZEELIB) $(EZEELIB).1: $(OBJECTS:%=$(OBJ_DIR)/%) $(LD) $(LFLAGS) $(OBJECTS:%=$(OBJ_DIR)/%) -o $(EZEELIB).1 $(OBJ_DIR)/%.o: %.cc $(CC) $(CFLAGS) -c $< -o $@ clean: rm -f $(OBJECTS:%=$(OBJ_DIR)/%) $(EZEELIB) $(EZEELIB).1 HERE ARE THE PROBLEM LINES: ============================ ezee.h: /* sizeof the image data header's records important for byte reversal */ size_t idcodefield = sizeof(int32); size_t dsexfield = sizeof(int16); size_t dtypefield = sizeof(EzeeType); size_t dcomplexfield = sizeof(int16); size_t nrowsfield = sizeof(int16); size_t ncolsfield = sizeof(int16); size_t nbandsfield = sizeof(int16); size_t nsetsfield = sizeof(int16); size_t data_sizefield = sizeof(ulong);
From: jrwst30@pitt.edu (James Wright) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 15:32:59 GMT Organization: Pitt Message-ID: <jrwst30-0509971132340001@jrwst30.j.resnet.pitt.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> i don't think that this is true. First of all, why would motorola do this?If they were worried about the falloff from cloners being missed(they only amount to about 300,000 machines), i think apple jumping ship would put a little more of a dent in their pockets since they are going to ship about 3 million machines this year. secondly, IBM relies on the Power Arhictecture for their workstation and supercomputing machines ; the next generation of these machines will contain massively parralel configurations of top of the line PowerPc's. . It just doesn't compute..lock this up in the rumor file. Lastly, if this is true, apple may as well stop producing machines now. Who in the hwll would want a machine that no one is going to produce software for. If apple discontinues powerpc, and switches to intel machines, which platfor do youthink the software is going to be written for?i know rhapsody is cross platform, but the apps aren't; they still must be recompiled for both intel and power pc. If this ever happened, apple would alienate alot of customers who just purcased powerpc macs, and would lose them to Microsoft simply because of stability.
From: randy.patton@vt.edu (Randy Patton) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 12:04:44 -0500 Organization: Virginia Tech Department of English Message-ID: <randy.patton-0509971204440001@news.vt.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit In article <B03540AE-CE02A@206.165.44.43>, "Lawson English" <english@primenet.com> wrote: > > What is really fun is that I'm starting to get details about what my "Very > Bad Thing" may turn out to be: > > Not Steve or Larry. They're after The NeXT BIg Thing, which is Net > Computers, with high-end servers. Hey, programmers: How hard would it be for someone to write an NC emulator for a desktop OS? I imagine myself in five years running a dual-boot BeOS/Linux Intel clone, with an inexpensive emulator package (Connectix Virtual NC?) to handle set-top and other NC functions. How feasible is this? Can we have our cake and eat it too?
From: "K. S. Brønnick" <broennic@online.no> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 16:02:52 GMT Organization: Telenor Online Public Access Message-ID: <01bcba15$447c2f00$48e04382@broennic.online.no> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> Lawson English <english@primenet.com> skrev i artikkelen <B03540AE-CE02A@206.165.44.43>... > Matthew Vaughan <see-below@not-my-address.com> said: > > > I think he meant that Ellison had originally attempted to buy Apple, and > > convert it into an NC manufacturer. Now he gets appointed to the board > > with essentially no money invested, and accomplishes the same thing. > > > > What is really fun is that I'm starting to get details about what my "Very > Bad Thing" may turn out to be: > > JObs was told by Motorola to stop playing games with clone-makers or face > losing PowerPC's from Motorola. Jobs replied that he didn't care since Net > Computers could just as easily run on x86 CPUs. (dare I point out that > Rhapsody is cross-platform?) > > Guess what? > > No more PowerPCs from Motorola starting in 1999. Word is that the decision > is moving through channels, even as we speak. > > Since Rhapsody is cross-platform, who cares if Apple doesn't make PowerMacs > anymore? > > Not Steve or Larry. They're after The NeXT BIg Thing, which is Net > Computers, with high-end servers. > > It all becomes clear now... > Maybe I'm wrong here- but aren't the PowerPC CPU's used in lots of other products than Macs- like in cars for instance? Regards K.S. Brønnick
From: "Mark Szamrej" <szamrej@no-spam.centerline_dot_com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 16:24:18 GMT Organization: Micron Electronics, Inc. Message-ID: <01bcba18$51e27bb0$a201ef8c@ubu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <randy.patton-0509971204440001@news.vt.edu> A NC is basically a "simple box" that runs a Web browser & Java. If you have a Web browser on your desktop system then you have a "smart" NC. - Mark - Randy Patton <randy.patton@vt.edu> wrote in article <randy.patton-0509971204440001@news.vt.edu>... : : Hey, programmers: : : How hard would it be for someone to write an NC emulator for a desktop OS? : : I imagine myself in five years running a dual-boot BeOS/Linux Intel clone, : with an inexpensive emulator package (Connectix Virtual NC?) to handle : set-top and other NC functions. : : How feasible is this? Can we have our cake and eat it too? :
From: acurylo@inmediapresents.nospam.com (Alex Curylo) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 09:59:58 -0700 Organization: InMedia Presentations Message-ID: <acurylo-0509970959590001@p2-29.van.tvs.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> In article <B03540AE-CE02A@206.165.44.43>, "Lawson English" <english@primenet.com> wrote: > JObs was told by Motorola to stop playing games with clone-makers or face > losing PowerPC's from Motorola. Jobs replied that he didn't care since Net > Computers could just as easily run on x86 CPUs. Well, well, well, well, well. It's nice to hear even a wild rumour that backs up exactly what I've been saying on the Knife boards for a while now. Apparently S. Jobs shares my conviction that PPC hw is going to have all the market appeal in six months of a NeXTStation post-NeXTStep/Intel... ---- Alex Curylo -- alex@witty.com
From: wildcat@ekx.infi.net Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 16:54:31 GMT Organization: InfiNet Message-ID: <wildcat-0509971255160001@news.infi.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <5up1dc$ndg$1@news.lth.se> In article <5up1dc$ndg$1@news.lth.se>, d89cb@efd.lth.se (Christian Brunschen) wrote: > Beginning of 1999 -- that's a little over a year away. By that time, > Rhapsody Unified will have been shipping for several months, and > Rhapsody Premier for almost a year. Most new application development > for the Mac platform, should be happening under Rhapsody -- which > means that, for new applications, indeed we do not care if we are > running on PowerPC or on Intel. > That's funny. All I've read about for the past two years is how superior the PowerPC chip is to intel. Now that there might not be any PowerPC chips, you say it really doesn't matter? Righttttt...... --SteveB--
From: acurylo@inmediapresents.nospam.com (Alex Curylo) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 10:02:54 -0700 Organization: InMedia Presentations Message-ID: <acurylo-0509971002540001@p2-29.van.tvs.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> <34100896.6928@A1.tch.harvard.edu> > I'd like to see a source for this rumor. Well, common sense and a little knowledge of NeXT's history should lead you to more or less this conclusion on your own. > I sincerely can't believe that > they would do this. Seems like the best business proposition they have available to me, and it fits nicely with the clone fracas and everything else that's been said about nixing CHRP etc. etc. etc. doesn't it now?? ---- Alex Curylo -- alex@witty.com
From: acurylo@inmediapresents.nospam.com (Alex Curylo) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 10:10:02 -0700 Organization: InMedia Presentations Message-ID: <acurylo-0509971010020001@p2-29.van.tvs.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> In article <01bcba15$447c2f00$48e04382@broennic.online.no>, "K. S. Brønnick" <broennic@online.no> wrote: > Maybe I'm wrong here- but aren't the PowerPC CPU's used in lots of other > products than Macs- like in cars for instance? The PowerPC family dominates a large number of embedded systems markets. Besides cars, there's a PPC in pretty much every piece of medical imaging hardware, for instance. Not to mention the Mars Rover. (PPC -- 100% of the Martian computing installed base! There's an ad for you huh?) See, in embedded markets you don't have idiot consumers who believe what the media and Intel ads tell them and make their decisions based on FUD, you generally have people whose interest it is in to make the most rationally cost-effective decision. Which, generally, would lead to one picking a PPC architecture chip. ---- Alex Curylo -- alex@witty.com
From: acurylo@inmediapresents.nospam.com (Alex Curylo) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 10:13:35 -0700 Organization: InMedia Presentations Message-ID: <acurylo-0509971013360001@p2-29.van.tvs.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <randy.patton-0509971204440001@news.vt.edu> In article <randy.patton-0509971204440001@news.vt.edu>, randy.patton@vt.edu (Randy Patton) wrote: > How hard would it be for someone to write an NC emulator for a desktop OS? You run a reasonably recent version of Netscape or MSIE, with Java enabled, yes? Then you have it now. ---- Alex Curylo -- alex@witty.com
From: ga@ed4u.com (G. Apple) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class (long) Date: 5 Sep 1997 17:58:01 GMT Organization: Advanced Communications Engineering, Inc. Message-ID: <5uph79$pvu$1@gte1.gte.net> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> <5umnvq$ju9$1@gte2.gte.net> <34100283.0@epflnews.epfl.ch> In article <34100283.0@epflnews.epfl.ch>, marco@sente.ch (Marco Scheurer) wrote: > So "+" has to be THE operator to translate a rectangle and moveTo: (or > translateBy:) would be less acceptable? You are confusing operator and > symbol. Anyway, you're wrong on both counts: mathematics and > programming. > > MATHEMATICS > In my practice of math, "+" is always used for a function with the > properties of addition in Z (the set of integers), and defines a group > with the set on which it operates. Generally (G, .) denotes the group > of the set G with the operation "." . > > Practically, . is replaced with + when it is commutative and shows > other analogies with the addition of real numbers: for instance the > addition of complex numbers, the addition of vectors. These analogies > can also, in the case of rings, refer to the properties of addition > with respect to multiplication, for instance the behavior of their > respective identity elements. > > Nevertheless, even without refering to these analogies, and not event > the commutativity of addition (is your overloaded + operator > commutative? I didn't think so), Of course it's commutative. Why would you think otherwise? Also, it's not "my" operator, it's ACS, derived from similar classes in MacApp which has often been credited as the best example of an application framework. It works very well and even allows easily moving between 16-bit and 32-bit coordinate systems. > The use of + you described is bad practice. What do you get? It is > just as much work to overload + as it is to define a new function, > such as "translate", or maybe "tr" or "t", which are symbols I've > found in geometry texts to denote a translation. You take risks > because it is ambiguous: it could very well be understood has a form > of union. You loose readability. > > I undertand that some people find overloading operators in general, > and + in particular, at the slightest reason, cute, or cool. > Unfortunately, these people often include programmers. > Ok, so maybe I missed the target at whom I was aiming. My real target was the chorus of objectors to any kind of operator overloading or multiple inheritance. It seems to me that most of these are in the Objective-C and Java camps, many of the latter particularly who have jumped on Java because they have heard it was easy and then believe "Now I are a programmer too, so don't confuse me by changing the meaning of a plus sign". Apparently, that shoe does not fit you, so I offer my apologies. (Like the cartoon says, on the Internet no one knows that you're really a dog. :-) Although I agree with your comments about "slightest reason, cute, or cool", I do not agree with that (properly used) operator overloading make things "ambiguous" or "lose readibility". For example, use of complex numbers and matrices in C++ can make an engineering algorithm vastly more understandable than a bunch of Fortran gobbledygook required to do the same thing. I have often carried principles of abstract algebra over to more classical engineering problems (not just the field of error correction coding) where I was able to get relatively simple solutions to what would have been very messy algorithms otherwise. (I recently got a large plaque from TRW (where I used to work) for a patent that issued on one of these. Operator overloading and multiple inheritance are two of the most powerful features of C++, and I just hate like hell to lose them in the stampede to Java and Objective C. -- G. Gordon Apple, PhD The Ed4U Project Advanced Communications Engineering, Inc. Redondo Beach, CA ga@ed4u.com
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 11:02:44 -0700 Organization: Netcom Message-ID: <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> In article <01bcba15$447c2f00$48e04382@broennic.online.no>, "K. S. Brønnick" <broennic@online.no> wrote: > Maybe I'm wrong here- but aren't the PowerPC CPU's used in lots of other > products than Macs- like in cars for instance? It's a pretty strong rumor at this point that all development (except bug fixing and yield gains) has been stopped on Macintosh related PowerPC's. Motorola will stop shipping them sometime in 1998 or early 1999. Of course, development of other PPC implementations, used in many Motorola products, will continue. Get ready for Steve's next big thing...a slotless, fanless, underpowered and overpriced computer with a little RAM, and even smaller HD (if one at all), and a permanent connection to a network where you wait for all your applications and data. Imagine! Everything that's yours stored somewhere else! No privacy! Web like wait times for your applications and files! A dead computer when the network goes down! It's the future people. A graphic designer won't mind waiting while his/her NC swaps MB after MB of program and file data back and forth over an Ethernet line to apply a single filter because MIS will save so much money on support costs. They will love NC's and jump right on them to save in support costs, just like they jumped right on Macintosh to save in support costs, right??? They will gladly dump all their hardware just to be able to run slow Java apps on a slow NC...right? Or dump native apps in favor of Java interpreted apps served over a bogged down Ethernet line, right? Whoops! Some repair guy just smashed through a wall and the networking cable. Everybody go home now.... And consumers will flock to this new technology. Imagine loading Word at 28.8K. Heck, why doesn't AOL jump on this? All they have to do is integrate a word processor, spreadsheet, and DB into the AOL client software. Don't all of you want to do your work via AOL's server and store your data there? If that picture doesn't send chills up and down your spine, send your resume to Apple. They need people like you with "vision". -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: Cosmo Roadkill <cosmo.roadkill%bofh.int%rauug.mil.wi.us@bofh.int> Newsgroups: comp.sys.next.programmer,uswest.test Subject: cmsg cancel <5upb4n$pif6@utoepia.advtech.uswest.com> Control: cancel <5upb4n$pif6@utoepia.advtech.uswest.com> Date: 05 Sep 1997 18:02:37 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5upb4n$pif6@utoepia.advtech.uswest.com> Sender: NNTP Component <nntpcomp@microsoft.com> My From: line has been fudged because many test newsgroup autoresponders respond to control messages. My apologies! Please see the X-Cancelled-By: line for my proper address. Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc From: nagle@netcom.com (John Nagle) Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Message-ID: <nagleEG1ro9.KCB@netcom.com> Organization: Netcom On-Line Services References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <randy.patton-0509971204440001@news.vt.edu> Date: Fri, 5 Sep 1997 18:07:21 GMT Sender: nagle@netcom6.netcom.com randy.patton@vt.edu (Randy Patton) writes: >How hard would it be for someone to write an NC emulator for a desktop OS? >I imagine myself in five years running a dual-boot BeOS/Linux Intel clone, >with an inexpensive emulator package (Connectix Virtual NC?) to handle >set-top and other NC functions. >How feasible is this? Can we have our cake and eat it too? You can buy something like that right now from QNX (www.qnx.com). It's a set-top-box program running on QNX. Runs on well on 386 machines, even. John Nagle
From: Cosmo Roadkill <cosmo.roadkill%bofh.int%rauug.mil.wi.us@bofh.int> Newsgroups: comp.sys.next.programmer,uswest.test Subject: cmsg cancel <5upb7k$pi711@utoepia.advtech.uswest.com> Control: cancel <5upb7k$pi711@utoepia.advtech.uswest.com> Date: 05 Sep 1997 18:10:48 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5upb7k$pi711@utoepia.advtech.uswest.com> Sender: NNTP Component <nntpcomp@microsoft.com> My From: line has been fudged because many test newsgroup autoresponders respond to control messages. My apologies! Please see the X-Cancelled-By: line for my proper address. Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
Date: Fri, 05 Sep 1997 13:35:13 -0600 From: David Stes <stes@can.nl> Subject: Re: Objective-C Fragile Base Class Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Message-ID: <873483955.2013@dejanews.com> Organization: Computer Algebra Netherlands References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <x7afhuyivf.fsf@tom.ics.ele.tue.nl> <jesjones-ya02408000R0309972014170001@news.halcyon.com> <x7en75cqmn.fsf@tom.ics.ele.tue.nl> <jesjones-ya02408000R0409972051430001@news.halcyon.com> In article <jesjones-ya02408000R0409972051430001@news.halcyon.com>, jesjones@halcyon.com (Jesse Jones) wrote: > > I heard at WWDC (or maybe MacHack) that Apple's > Objective-C is going to use the C++ exception model. A plan that has been around for a while (since around '87), and that I've implemented to try out, is to use Objective-C Blocks for this. It makes it easy to replace error messages by exceptions: - at:(unsigned)i { if (i > size) [self error:"index %i > size %i",i,size]; becomes, - at:(unsigned)i { if (i > size) [{[self error:"index %i > size %i",i,size];} raise]; The stuff inside {} is a Block and can refer to "self", "i" and "size" (where "size" could be an instance or class variable). If the exception is *uncaught*, the two examples behave virtually identical. Specifically: the stack backtrace towards the method that was raising the exception (the method -at:) is *NOT* lost. This means that you (in a debugger) don't have to set extra breakpoints to find out where the exception comes from. David. -------------------==== Posted via Deja News ====----------------------- http://www.dejanews.com/ Search, Read, Post to Usenet
From: d89cb@efd.lth.se (Christian Brunschen) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 18:37:17 GMT Organization: Lund Institute of Technology, Sweden Message-ID: <5upjgt$sme$1@news.lth.se> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> NNTP-Posting-User: d89cb In article <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com>, Daniel L. Taylor <LTaylor7@ix.netcom.com> wrote: > >Get ready for Steve's next big thing...a slotless, fanless, underpowered >and overpriced computer with a little RAM, and even smaller HD (if one at >all), and a permanent connection to a network where you wait for all your >applications and data. I wonder where this rumor comes from. Where has Steve Jobs said anything about stripping machines of powerful CPU, Memory, secondary storage ? Ah, I get it -- the letters 'NC' scare you. Well, a 'NC' need not be slow, RAMless, or diskless. I beleive Steve Jobs' "vision of an NC" includes _Rhapsody_. Rhapsody needs, among other things: a) a _fast_ CPU b) _lots_ of RAM c) dynamic libraries, applications, etc -- quickly accessible. From, for instance, a hard disk. > >Imagine! Everything that's yours stored somewhere else! No privacy! Web >like wait times for your applications and files! A dead computer when the >network goes down! Oh Please. Lots of people live quite comfortably with having their home directories mounted over a LAN, using, say, NFS (under Unix). If you start putting your servers and clients far apart, connected through 28k8 modems, of course performancs will suffer. But put them on a 100Mb/s Ethernet, and I doubt anyone will have any complaints -- especially since people will be able to choose any one of a bunch of identical computers and get their own personal computing environment. > >It's the future people. A graphic designer won't mind waiting while >his/her NC swaps MB after MB of program and file data back and forth over >an Ethernet line to apply a single filter because MIS will save so much >money on support costs. They will love NC's and jump right on them to save >in support costs, just like they jumped right on Macintosh to save in >support costs, right??? It's about adminitering all machines centrally. If all machines are NC:s - in the sense that they aren't 'individual' machines, but rather 'clones' regarding OS, software, and even most hardware - then you can shave off all support costs regarding support of the 'individual' machines, since all machines will be interchangable. Plus, you will be able to have a few 'fresh' machines in 'reserve', so if any machine in the 'field' breaks down, you can swap it for one that works, without further interruption to the work of whoever is using the machine in question. > >They will gladly dump all their hardware just to be able to run slow Java >apps on a slow NC...right? Or dump native apps in favor of Java >interpreted apps served over a bogged down Ethernet line, right? Whoops! >Some repair guy just smashed through a wall and the networking cable. >Everybody go home now.... Um, Rhapsody runs, first and foremost, compiled Objective-C applications. JIT-compiled Java is hardly 'interpreted' either, btw. Also, how 'bogged- down' your Ethernet cable is, largely depends on other things - applications tend to get loaded once, and perhaps load more data (libraries, bundles) every now and then. Plus of course, you usually work with some sort of document, which tends to get loaded, saved (often auto-saved to minimize loss in case the application crashes). Still, on a 100MB/s Ethernet, this would hardly 'bog down' the net. If you strip your machines of RAM and swap over the net, then yes -- but noone has suggested that! > >And consumers will flock to this new technology. Imagine loading Word at >28.8K. Heck, why doesn't AOL jump on this? All they have to do is >integrate a word processor, spreadsheet, and DB into the AOL client >software. Don't all of you want to do your work via AOL's server and store >your data there? If that picture doesn't send chills up and down your >spine, send your resume to Apple. They need people like you with "vision". You have heard the expression 'FUD', perhaps? Stands for 'Fear, Uncertainty, and Doubt'. This is usually backed up with baseless assertions, sometimes even outright lies. Just like you have been writing. Thank you for your cooperation. The computer is your friend. Happiness is mandatory. Please report to the nearest termination center immediately. > >-- >Daniel L. Taylor >LTaylor7@ix.netcom.com >================================================================= >The Heritage Foundation /// Town Hall >http://www.heritage.org/heritage/ /// http://www.townhall.com/ >================================================================= // Christian Brunschen
From: herren@flannet.middlebury.edu (David Herren) Newsgroups: comp.sys.next.programmer Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 5 Sep 1997 14:29:30 -0400 Organization: Language Schools of Middlebury College Message-ID: <msg137882.thr-465900b1.54c5638@flannet.middlebury.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> Mime-Version: 1.0 Content-Type: text/enriched; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Content-ID: <msg137882.thr-465900b1.54c5638.part0@flannet.middlebury.edu> <bold>broennic@online.no,UseNet writes:</bold> >Maybe I'm wrong here- but aren't the PowerPC CPU's used in lots of other= >products than Macs- like in cars for instance? And the Martian Rover... -- = David D. Herren www.cet.middlebury.edu/herren Assoc. Dir. for Tech. & Instruction herren@flannet.middlebury.edu Center for Educational Technology voice: (802)443-5746 Middlebury College, Middlebury, VT 05753 fax: (802)443-2053
From: Tony Invencio <"tonyinv"@hotmail .dot com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 14:55:15 -0400 Organization: Have None Message-ID: <5upkik$45m@dfw-ixnews10.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Daniel L. Taylor wrote: > > In article <01bcba15$447c2f00$48e04382@broennic.online.no>, "K. S. > Brønnick" <broennic@online.no> wrote: > > > Maybe I'm wrong here- but aren't the PowerPC CPU's used in lots of > other > > products than Macs- like in cars for instance? > > It's a pretty strong rumor at this point that all development (except > bug > fixing and yield gains) has been stopped on Macintosh related > PowerPC's. > Motorola will stop shipping them sometime in 1998 or early 1999. > > Of course, development of other PPC implementations, used in many > Motorola > products, will continue. > > Get ready for Steve's next big thing...a slotless, fanless, > underpowered > and overpriced computer with a little RAM, and even smaller HD (if one > at > all), and a permanent connection to a network where you wait for all > your > applications and data. > > Imagine! Everything that's yours stored somewhere else! No privacy! > Web > like wait times for your applications and files! A dead computer when > the > network goes down! > > It's the future people. A graphic designer won't mind waiting while > his/her NC swaps MB after MB of program and file data back and forth > over > an Ethernet line to apply a single filter because MIS will save so > much > money on support costs. They will love NC's and jump right on them to > save > in support costs, just like they jumped right on Macintosh to save in > support costs, right??? > > They will gladly dump all their hardware just to be able to run slow > Java > apps on a slow NC...right? Or dump native apps in favor of Java > interpreted apps served over a bogged down Ethernet line, right? > Whoops! > Some repair guy just smashed through a wall and the networking cable. > Everybody go home now.... > > And consumers will flock to this new technology. Imagine loading Word > at > 28.8K. Heck, why doesn't AOL jump on this? All they have to do is > integrate a word processor, spreadsheet, and DB into the AOL client > software. Don't all of you want to do your work via AOL's server and > store > your data there? If that picture doesn't send chills up and down your > spine, send your resume to Apple. They need people like you with > "vision". > I was at the Borland Developers Conference in Nashville this summer and Ellison gave one of the Keynotes (via sattelite) on this very subject. He was of course spewing an anti-MS Pro-Oracle "solution". When we got to Q&A someone made your very point. And strangly enough Ellison mentioned AOL, the crowd audibly gasped. Then he went into some bullshit about the home having multiple "Compuliances" (I just made that up!) connected to a _server_ right there in the home. It was pretty funny cause as he was saying it he started to trail off cause he knew just how crazy it was. It was obvious that this crowd of MIS/IT people and Developers was very skeptical. Whats amusing is that everyone talks about Rhapsody but is forgetting Oracle's roll in this. I'm not sure where it might fit but Ellision has an agenda and it explicitly includes makeing more money for _HIS_ company. A healthy Apple may not include a MacOS afterall.... -- Tony Invencio Strange Solutions Software "The fault, dear Brutus, is not in our stars, But in ourselves, that we are underlings." Cassius
From: mpaque.spa-am@nospam.wco.com (Mike Paquette) Newsgroups: comp.sys.next.programmer Subject: Re: Automatic spinning cursor of Mach not going away ?? Date: Fri, 05 Sep 1997 19:20:34 GMT Organization: Electronics Service Unit No. 16 Message-ID: <34104142.10728997@news.wco.com> References: <5uje9s$1q2$1@concorde.ctp.com> On 3 Sep 1997 10:31:24 GMT, tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) wrote: >But for one of my applications, under certain yet unknown circumstances, it >starts spinning...and never stops. >The application is fully active and works just fine..but the cursor is >spinning and I have not found an obvious way to reset it. > > [[NSCursor arrowCursor] set] >does not cure the disease :-( This sets the user's cursor, and doesn't impact the wait cursor. The wait cursor normally appears when your application has failed to consume any events posted to it for a period of time (roughly a second by default). Apps should try to avoid getting their event processing thread tied up for too long. Of course, there are always those special occasions where you just can't help it. Your app can live with the wait cursor, which gives the user a clue that the app is busy, or it can intervene through the NSDPSContext instance methods setWaitCursorEnabled: and isWaitCursorEnabled. If you believe that your app is responding to events, and you are still getting the wait cursor, you might want to look really hard at your event processing, rather than blithly disabling the wait cursor.
From: tzs@halcyon.com (Tim Smith) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 12:28:00 -0700 Organization: Northwest Nexus Inc. Message-ID: <5upmg0$oun$1@halcyon.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> Lawson English <english@primenet.com> wrote: >JObs was told by Motorola to stop playing games with clone-makers or face >losing PowerPC's from Motorola. Jobs replied that he didn't care since Net That is extremely unlikely. Almost all large companies that are involved in more than one line of business let the separate business units make mostly independent decisions. Unless the Motorola group that makes the PPC chip is the same as the group that makes Mac clones (is it?), the chip group probably couldn't care less if Apple screws the clone group, as long as the chip group sells a lot of chips. --Tim Smith
From: cdstrand@onramp.net (Dave) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 15:11:31 -0500 Organization: Could be better Message-ID: <cdstrand-0509971511310001@cpsidfw.flash.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> In article <5upjgt$sme$1@news.lth.se>, d89cb@efd.lth.se (Christian Brunschen) wrote: (( cuts )) > I beleive Steve Jobs' "vision of an NC" includes _Rhapsody_. Rhapsody needs, > among other things: > a) a _fast_ CPU > b) _lots_ of RAM > c) dynamic libraries, applications, etc -- quickly accessible. From, for > instance, a hard disk You may be right, though as I read it, NC means some sort of shell with all of the programs and storage on the server. Otherwise what's the point? Actually the speed of the CPU on a NC would actually be irrelevant, and the only important thing would be the speed of the interface between the NC and the server to which it is networked. (( cuts )) -- Dave cdstrand@onramp.net
From: "John Huang" <john.nospam@scenarist.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 20:22:56 GMT Organization: North Bay Network, Inc. news server - not responsible for content Message-ID: <01bcba3a$8cbc6e40$b55db8ce@paper> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> <341168EE.61A1@ix.netcom.com> > > Talk of using an x86 for any Apple product or MacOS machine (as > opposed to Rhapsody) truly disgusts me. The PowerPC is a superior chip, > is more modern and Apple should run its platform on this chip (it gives > MAc superior performance in graphics) and Newtons on StrongARMs. We > expect the best out of Apple and AIM, all players, must get their act > together and not talk of killing this chip. The talk should be of > continuing to outpace the x86 and, I guess, the vapour Merced. Not this. > If Steve stands for quality and pushing the enveloppe this is truly > unacceptable. > > Ben Smith Wake up guys, business world is not child play. If Motorola stops making PPC, it will because there isn't enough volumn demand in the market to make a profit. And it truely doesn't make sense to have both IBM and Motorola plants, each of them only running at one third capacity and losing money. John Next time I turn around, I bet GM cancel some car model and someone will somehow blame it on Apple
From: don_arb@wolfenet.com (Don Arbow) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 12:14:21 -0700 Organization: None Message-ID: <don_arb-0509971214210001@sea-ts1-p59.wolfenet.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> In article <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com>, LTaylor7@ix.netcom.com (Daniel L. Taylor) wrote: : In article <01bcba15$447c2f00$48e04382@broennic.online.no>, "K. S. : Brønnick" <broennic@online.no> wrote: : : > Maybe I'm wrong here- but aren't the PowerPC CPU's used in lots of other : > products than Macs- like in cars for instance? : : It's a pretty strong rumor at this point that all development (except bug : fixing and yield gains) has been stopped on Macintosh related PowerPC's. : Motorola will stop shipping them sometime in 1998 or early 1999. : : Of course, development of other PPC implementations, used in many Motorola : products, will continue. : : Get ready for Steve's next big thing...a slotless, fanless, underpowered : and overpriced computer with a little RAM, and even smaller HD (if one at : all), and a permanent connection to a network where you wait for all your : applications and data. : : Imagine! Everything that's yours stored somewhere else! No privacy! Web : like wait times for your applications and files! A dead computer when the : network goes down! : : It's the future people. A graphic designer won't mind waiting while : his/her NC swaps MB after MB of program and file data back and forth over : an Ethernet line to apply a single filter because MIS will save so much : money on support costs. They will love NC's and jump right on them to save : in support costs, just like they jumped right on Macintosh to save in : support costs, right??? : : They will gladly dump all their hardware just to be able to run slow Java : apps on a slow NC...right? Or dump native apps in favor of Java : interpreted apps served over a bogged down Ethernet line, right? Whoops! : Some repair guy just smashed through a wall and the networking cable. : Everybody go home now.... : : And consumers will flock to this new technology. Imagine loading Word at : 28.8K. Heck, why doesn't AOL jump on this? All they have to do is : integrate a word processor, spreadsheet, and DB into the AOL client : software. Don't all of you want to do your work via AOL's server and store : your data there? If that picture doesn't send chills up and down your : spine, send your resume to Apple. They need people like you with "vision". : Most of the hype about NC's deals with corporations and large intranets where employees are using $2500 PC workstations to play solitaire. The main push behind using NC's is that these companies can get rid of those pricy machines, if the users are only going to run a word processor and a database client. Centralized storage for programs and data makes sense for maintenance and administration costs. No one has suggested that NCs will replace the "personal" market. As for storing of personal data on a remote server, quite a few people do this very thing every day. I have 5 megabytes of storage available on my ISP's computer, no one but myself can access it. Beware of the FUD... Don -- Don Arbow, Partner, CTO EveryDay Objects, Inc. don_arb@wolfenet.com <-- remove underscore to reply http://www.edo-inc.com
From: mmalcolm crawford <malcolm@plsys.co.uk> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 20:38:01 GMT Organization: P&L Systems Message-ID: <5upqj9$f1$1@ironhorse.plsys.co.uk> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <acurylo-0509970959590001@p2-29.van.tvs.net> In-Reply-To: <acurylo-0509970959590001@p2-29.van.tvs.net> On 09/05/97, Alex Curylo wrote: >Apparently S. Jobs shares my conviction that PPC hw is going to have all >the market appeal in six months of a NeXTStation post-NeXTStep/Intel... > Actually, NeXTstations are still (four years after NeXT stopped making them) quite popular; we had another order for one today. Best wishes, mmalc.
From: Jonathan Harker <Limeyville@spamtoNull.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 16:55:02 -0400 Organization: Netcom Message-ID: <341071A6.ACCCB72F@spamtoNull.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Daniel L. Taylor wrote: > > > Can't somebody sue him? Apple customers? Stockholders? Motorola? Can't > somebody drive him out before he kills Apple, the Mac, and Rhapsody? The sad thing is that the stockholders aren't savvy enough about the industry as a whole to see and understand the ruination Jobs is heralding. For some weird reason, they think he's qualified to call the shots in the company. The fact that NeXT has been an abysmal market failure seems all but lost on them, as does the fact that Jobs hasn't been an active Apple advocate for more than a decade. Well, maybe they just feel he's better than nothing....Big mistake, IMHO. -- Jonathan Hollingsworth Harker New astronomical data shows the age of the universe as measured by the latest galactic red-shift data at between 8 and 12 billion years. However, hydrogen decay data proves indisputably that many stars are 15-20 billion years of age. Modern astronomical theory about the formation and age of the universe is so accurate that it tells us that many stars are older than the universe in which they exist. And you laughed when they said the world was flat?
From: Jonathan Harker <Limeyville@spamtoNull.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 17:09:05 -0400 Organization: Netcom Message-ID: <341074F1.8F26E657@spamtoNull.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <5up1dc$ndg$1@news.lth.se> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Christian Brunschen wrote: > > Beginning of 1999 -- that's a little over a year away. By that time, > Rhapsody Unified will have been shipping for several months, and > Rhapsody Premier for almost a year. Most new application development > for the Mac platform, should be happening under Rhapsody -- which > means that, for new applications, indeed we do not care if we are > running on PowerPC or on Intel. > > > Rhapsody's cross-platform abilities mean that Apple is not 'locked' > into the PowerPC processor anymore, and can act freely in regard to > even someone like Motorola, their chip supplier for PowerPC. 'Freely' > meaning, of course, 'in Apple's self-interest', as defined by Apple's > leadership. Yea, this sounds so nifty in theory. It *is* a fine concept, no doubt about it. The central problem here, the fatal flaw, though, is that it *completely ignores the character and the identity of the traditional Apple market.* The "core identity" of the Apple market is that it *exists* as an alternative to the Wintel market. It has always been so. Nine out of every ten Apple customers will tell you this is why they use, and are loyal to, Apple Computer. Many of them fully believe that the PPC architecture is superior to Intel architecture, and so they make a conscious choice to buy Mac, instead. If PPC is dead, and the Mac OS for the PPC is dead (Rhapsody for PPC), then Apple computer has just removed the last barriers to its traditional customer base buying something *other than Apple.* The very last barriers will drop at that time. The "Mac" will cease to have a cohesive identity, and will melt away into the backdrop of the enormous Wintel landscape. Apple will certainly die, taking the Mac along with it. Rhapsody will merely be absorbed and digested. How long has NeXT for Intel existed as a marketable product? How well has it done? There is no reason whatever to think that a Rhapsody for Intel, devoid of the PPC and traditional Apple identity, will do any better than NeXT, which had neither of those things, either. -- Jonathan Hollingsworth Harker New astronomical data shows the age of the universe as measured by the latest galactic red-shift data at between 8 and 12 billion years. However, hydrogen decay data proves indisputably that many stars are 15-20 billion years of age. Modern astronomical theory about the formation and age of the universe is so accurate that it tells us that many stars are older than the universe in which they exist. And you laughed when they said the world was flat?
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 14:14:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B035C945-246BF@206.165.44.93> References: <5upmg0$oun$1@halcyon.com> nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.system, nntp://news.primenet.com/comp.sys.mac.misc MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Tim Smith <tzs@halcyon.com> said: > That is extremely unlikely. Almost all large companies that are involved > in more than one line of business let the separate business units make > mostly independent decisions. Unless the Motorola group that makes the > PPC chip is the same as the group that makes Mac clones (is it?), the > chip group probably couldn't care less if Apple screws the clone group, > as long as the chip group sells a lot of chips. My understanding is that the decision came from the very top of Motorola, just as Jobs is the "very top" of Apple. ------------------------------------------------------------------------- Build a better mousetrap, yes, but then market it like crazy, or instead of catching mice you'll only collect dust. -David Yeargin -------------------------------------------------------------------------
From: "Benjamin Smith" <benjs@ix.netcom.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 97 17:11:43 -0500 Organization: Netcom Message-ID: <B035EDF0-18AAE@207.92.153.95> References: <5upkik$45m@dfw-ixnews10.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit nntp://nntp.ix.netcom.com/comp.sys.next.advocacy, nntp://nntp.ix.netcom.com/comp.sys.mac.programmer.misc, nntp://nntp.ix.netcom.com/comp.sys.next.programmer, nntp://nntp.ix.netcom.com/comp.sys.mac.system, nntp://nntp.ix.netcom.com/comp.sys.mac.misc On Fri, Sep 5, 1997 1:55 PM, Tony Invencio <mailto:"tonyinv"@hotmail .dot com> wrote: >A healthy Apple may not include a MacOS afterall.... This is getting ridiculous. Remember in Boston, just a few months ago how Jobs spoke up the MacOS and then rampant silliness about Jobs not being interested in developing Rhapsody because he did not mention it. Isn't MacOS 8 selling like crazy and the plans for Allegro are full steam ahead. Ellison has said over and over he is interested in Big Macs, the current ones that run Mac OS and will run Rhapsody, and little Macs, the NCs for children, which Apple, buy the way, is making. People are going off of the deep end. Get a grip, folks. Ben Smith
From: mmalcolm crawford <malcolm@plsys.co.uk> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 21:08:00 GMT Organization: P&L Systems Message-ID: <5upsbg$10h$1@ironhorse.plsys.co.uk> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <cdstrand-0509971511310001@cpsidfw.flash.net> In-Reply-To: <cdstrand-0509971511310001@cpsidfw.flash.net> On 09/05/97, Dave wrote: >In article <5upjgt$sme$1@news.lth.se>, d89cb@efd.lth.se (Christian >Brunschen) wrote: > >(( cuts )) > >> I beleive Steve Jobs' "vision of an NC" includes _Rhapsody_. Rhapsody needs, >> among other things: >> a) a _fast_ CPU >> b) _lots_ of RAM >> c) dynamic libraries, applications, etc -- quickly accessible. From, for >> instance, a hard disk > >You may be right, though as I read it, NC means some sort of shell with >all of the programs and storage on the server. > If by "shell" you include "has the OS on a local hard drive", then I'm using an NC right now... NeXT made it in 1992. Others have made similar devices for a long time. Does anyone know what company produces Mogadon? I want to buy some of their shares -- I can see product sales increasing rapidly real soon now. Best wishes, mmalc.
From: cdstrand@onramp.net (Dave) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 16:36:43 -0500 Organization: Could be better Message-ID: <cdstrand-0509971636430001@cpsidfw.flash.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <cdstrand-0509971511310001@cpsidfw.flash.net> <5upsbg$10h$1@ironhorse.plsys.co.uk> In article <5upsbg$10h$1@ironhorse.plsys.co.uk>, mmalcolm crawford <malcolm@plsys.co.uk> wrote: (( cuts )) > If by "shell" you include "has the OS on a local hard drive", then I'm using > an NC right now... NeXT made it in 1992. Others have made similar devices > for a long time. Well, there you go. You're happy and, I assume, you have one at home and are quite happy with the performance of it there. (( cuts )) -- Dave cdstrand@onramp.net
From: schoinas@fox.cs.wisc.edu (Ioannis Schoinas) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 21:58:04 GMT Organization: University of WI, Madison -- Computer Sciences Dept. Message-ID: <5upv9c$6f5@spool.cs.wisc.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upkik$45m@dfw-ixnews10.ix.netcom.com> Tony Invencio <"tonyinv"@hotmail .dot com> writes: +Then he went into some bullshit about the home having multiple +"Compuliances" (I just made that up!) connected to a _server_ right +there in the home. It was pretty funny cause as he was saying it he +started to trail off cause he knew just how crazy it was. Actually, when I can afford it, this is what I have in mind as a hobby project. The technology is already available out there. Use old notebooks as java terminals, a fast ethernet around the house, a wireless network for things you want to carry with you, a big server, a cable modem to reach out (or whatever comes out of the telcos), a computer with all the 3d hardware with the TV. Yannis
From: Jonathan Harker <Limeyville@spamtoNull.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: Fri, 05 Sep 1997 18:12:19 -0400 Organization: Netcom Message-ID: <341083C3.B1F2D7@spamtoNull.com> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> <01bcb967$7a41bd70$b55db8ce@paper> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit John Huang wrote: > > > > First, nobody knows what the fee structure really is. If i remembered > correctly, 2 to 10 times (that means 100 to 500 dollars) per license was > What Apple thought was fair, and demanded, but the agreement was much > lower.....50 to 200 dollars per license, i remember reading somewhere. That's because your are continuously and consistently wrong when you use your $50 figure. That figure covers *only* Mac OS 7.6. The cloners paid an additional fee of $150, on average, for *hardware licenses on Apple designs.* The total average fee, then, under the FIRST AGREEMENT was somewhere around $200 per machine. I've seen figures as high $1500 that Apple insisted on in the NEW AGREEMENT for some high-end SERVER configurations the cloners were trying to build. -- Jonathan Hollingsworth Harker New astronomical data shows the age of the universe as measured by the latest galactic red-shift data at between 8 and 12 billion years. However, hydrogen decay data proves indisputably that many stars are 15-20 billion years of age. Modern astronomical theory about the formation and age of the universe is so accurate that it tells us that many stars are older than the universe in which they exist. And you laughed when they said the world was flat?
From: sanguish@digifix.com (Scott Anguish) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 22:09:59 GMT Organization: Digital Fix Development Message-ID: <5upvvn$j8h$1@news.digifix.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> <wildcat-0509971252300001@news.infi.net> In-Reply-To: <wildcat-0509971252300001@news.infi.net> On 09/05/97, wildcat@ekx.infi.net wrote: > >In article <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com>, >LTaylor7@ix.netcom.com (Daniel L. Taylor) wrote: > > >> >> Can't somebody sue him? Apple customers? Stockholders? Motorola? Can't >> somebody drive him out before he kills Apple, the Mac, and Rhapsody? >> >Can someone explain to me how a guy who now owns ONE share of Apple stock >has such absolute power over the company? WHO gave him such authority and >how can it be RESCINDED? > Considering that the stock has held its own during all this bullshit over the past few days, that CALPERS which owns 700,000 shares, seems confident, I DOUBT that the shareholders would be trying to kick him out. Sometimes hard times call for hard choices. But of course, all your worrying is based on RUMOR. -- Scott Anguish <sanguish@digifix.com> NEXTSTEP/OpenStep Information <URL:http://www.stepwise.com>
From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Newsgroups: comp.sys.next.programmer Subject: Re: Automatic spinning cursor of Mach not going away ?? Date: 5 Sep 1997 22:27:14 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5uq102$91q$1@concorde.ctp.com> References: <5uje9s$1q2$1@concorde.ctp.com> <34104142.10728997@news.wco.com> mpaque.spa-am@nospam.wco.com (Mike Paquette) wrote: >Of course, there are always those special occasions where you just >can't help it. Your app can live with the wait cursor, which gives >the user a clue that the app is busy, or it can intervene through the >NSDPSContext instance methods setWaitCursorEnabled: and >isWaitCursorEnabled. Thanx for the hint. I stumbled over those PS functions just the other day after my posting... >If you believe that your app is responding to events, and you are >still getting the wait cursor, you might want to look really hard at >your event processing, rather than blithly disabling the wait cursor. > ... but while I can "cure" the disease (get the normal arrow back again) with these calls, the real problem is somewhere in the AppKit and somehow leaves the redraws and propably some event and DPS stuff in a wierd state. In order to get a good responsiveness out of TheShelf I was making my [NSWorkspace lauchThisDoThatTakeYourTime] request in a separate thread. Even if the AppKit is not thread save this works quite nicely...but somehow sometimes it just catches the DPS thing on the wrong foot...and then it is spinning time. Reseting the cursor does the trick...but from that moment on regular NSView redraws behave a little strange. So the current workaround is the old performSelector:withObject:afterDelay:. But this means NSWorkspace blocks :-( aloha Tomi P.S: Ok...there are other possible no-blocking solutions as well...but all of them are more work then I'm willing to invest right now.
From: mmalcolm crawford <malcolm@plsys.co.uk> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 22:31:34 GMT Organization: P&L Systems Message-ID: <5uq186$2dl$1@ironhorse.plsys.co.uk> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <cdstrand-0509971511310001@cpsidfw.flash.net> <5upsbg$10h$1@ironhorse.plsys.co.uk> <cdstrand-0509971636430001@cpsidfw.flash.net> In-Reply-To: <cdstrand-0509971636430001@cpsidfw.flash.net> On 09/05/97, Dave wrote: >In article <5upsbg$10h$1@ironhorse.plsys.co.uk>, mmalcolm crawford ><malcolm@plsys.co.uk> wrote: > >> If by "shell" you include "has the OS on a local hard drive", then >> [by Dave's definition] I'm using an NC right now... NeXT made it in 1992. >> Others have made similar devices for a long time. > >Well, there you go. You're happy and, I assume, you have one at home and >are quite happy with the performance of it there. > Actually, no, I'm using it at work. It's still a fine machine which today -- rather unusually, I don't do much programming work any more -- I've been using to refine a "toy" app which will, I believe, soon run on Rhapsody. The real developers in the office are using rather faster machines for serious work, but if I wanted I could use my NeXTstation to work on a version of our spreadsheet, Mesa, which I trust would also compile on a Rhapsody system. Best wishes, mmalc. Malcolm Crawford (NeXTmail) malcolm@plsys.co.uk Tel: +44 1494 432422 P & L Systems Fax: +44 1494 432478 http://www.plsys.co.uk/~malcolm
From: pohl@xxx.screaming.org (Pohl Longsine) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 5 Sep 1997 22:57:13 GMT Organization: National Security Agency Message-ID: <slrn6114hi.2e2.pohl@xxx.screaming.org> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> <wildcat-0509971252300001@news.infi.net> <5upvvn$j8h$1@news.digifix.com> Scott Anguish <sanguish@digifix.com> wrote: >On 09/05/97, wildcat@ekx.infi.net wrote: >>> >>> Can't somebody sue him? Apple customers? Stockholders? >>>Motorola? Can't somebody drive him out before he kills Apple, >>>the Mac, and Rhapsody? >>> >>Can someone explain to me how a guy who now owns ONE share of >>Apple stock has such absolute power over the company? WHO >>gave him such authority and how can it be RESCINDED? Wow, there are some really gullible people in this thread. (Scott Anguish not included. :-) That's really funny: Steve Jobs the puppet-master, with Larry Ellison & Co. dangling on tiny little strings. Get real. >Considering that the stock has held its own during all this >bullshit over the past few days, that CALPERS which owns 700,000 >shares, seems confident, I DOUBT that the shareholders would >be trying to kick him out. Agreed. Did you see how Wired reported this? It was sooo slanted. It went something like "investor reactions were lukewarm -- I mean, like, the stock didn't jump *nearly* as much as it did when M$ invested $150M . . . fer sure." (That quote's not exact, but that was the spirit of it. ;-)
From: "Takechan" <take@oatmeal.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 6 Sep 1997 09:22:03 +0900 Organization: Takechan Press Message-ID: <5uq7nh$7al$1@nn-tk001.ocn.ad.jp> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <acurylo-0509971010020001@p2-29.van.tvs.net> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-2022-jp" Content-Transfer-Encoding: 7bit Alex Curylo wrote, [snip false information about PPC in Mars mission, etc.] >See, in embedded markets you don't have idiot consumers who believe what >the media and Intel ads tell them and make their decisions based on FUD, >you generally have people whose interest it is in to make the most >rationally cost-effective decision. Which, generally, would lead to one >picking a PPC architecture chip. People would be happy to use the PPC chip if there were a decent OS for it that didn't already run on Intel, and if the performance advantages justified the cost of the move and the reduction in software availability. Barring that, it still makes more sense to go with the industry standard. People are not being idiots as you suggest, but merely practical.
Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc From: Peter Spomer <pjsa0@juts.ccc.amdahl.com> Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Content-Type: text/plain; charset=us-ascii Message-ID: <3410ACD4.5A14@juts.ccc.amdahl.com> Sender: netnews@ccc.amdahl.com (Usenet Administration) Content-Transfer-Encoding: 7bit Organization: Amdahl Integrated Desktop Support References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> Mime-Version: 1.0 Date: Sat, 6 Sep 1997 01:07:33 GMT Lawson English wrote: > What is really fun is that I'm starting to get details about what my "Very > Bad Thing" may turn out to be: > > JObs was told by Motorola to stop playing games with clone-makers or face > losing PowerPC's from Motorola. Jobs replied that he didn't care since Net > Computers could just as easily run on x86 CPUs. (dare I point out that > Rhapsody is cross-platform?) > > Guess what? > > No more PowerPCs from Motorola starting in 1999. Word is that the decision > is moving through channels, even as we speak. That might make sense if Apple was the only customer for PowerPC chips. Maybe you're confusing this with Motorola's decision to stop work on CHRP?
From: Tony Invencio <"tonyinv"@hotmail .dot com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 21:40:35 -0400 Organization: Have None Message-ID: <5uqcak$kr@dfw-ixnews6.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upkik$45m@dfw-ixnews10.ix.netcom.com> <5upv9c$6f5@spool.cs.wisc.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Ioannis Schoinas wrote: > > Tony Invencio <"tonyinv"@hotmail .dot com> writes: > > +Then he went into some bullshit about the home having multiple > +"Compuliances" (I just made that up!) connected to a _server_ right > +there in the home. It was pretty funny cause as he was saying it he > +started to trail off cause he knew just how crazy it was. > > Actually, when I can afford it, this is what I have in mind as a hobby > project. The technology is already available out there. Use old > notebooks > as java terminals, a fast ethernet around the house, a wireless > network > for things you want to carry with you, a big server, a cable modem to > reach out (or whatever comes out of the telcos), a computer with all > the 3d hardware with the TV. > > Yannis Yes, but how practicle is that for "the rest of us" -- Tony Invencio Strange Solutions Software "The fault, dear Brutus, is not in our stars, But in ourselves, that we are underlings." Cassius
From: mschemen@concentric.net (Mathew Schemenaur) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: 6 Sep 1997 01:38:02 GMT Organization: Concentric Internet Services Message-ID: <mschemen-0409972046570001@ts005d05.ftw-tx.concentric.net> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> > > "To buttress their argument, they noted that the licensees have been paying > only about $50 per system for their Mac OS licenses. They neglected to > mention the separate hardware license fees the cloners have also been > paying -- in some cases, more than $180 per unit. (Note that most of this > money goes straight to Apple's bottom line, since there's no cost of goods > nor any new R&D expenditure required.) If this is true then every sell of the MacOS goes straight to Apple's bottom line. Forget about the huge investment in the original designs, the PowerPC conversion, PCI switch over, ... > But the biggest flaw in Apple's case is that it is based on the original > licensing contracts; somehow company spokespeople seem to have forgotten > that the clone vendors already agreed to new and very different terms. > These agreements were hammered out in early June, and by all accounts they > provided for significant increases in license fees -- a two- to 10-fold > increase, depending on the model, relative to previous combined hardware > and OS royalties, according to two major licensees. " Of course they are going to compare it to the legal contract and not an agreement that was never finalized by a CEO that is no longer with Apple. That does not make it a lie. > That's a new fee structure of $100 to OVER FIVE HUNDRED DOLLARS per box, if > Henry is to be trusted, that the clone-makers ALREADY had agreed to, > according to Gil Amelio in a public interview. According to the expert Harry Norr, Apple was already getting 100 to 280 per box when you concider software and hardware fees. So a jump to 100 to 500 does not seem like a jump at all. I think Gil got fired because he agreed to terms that were not much better then the original terms. At least not enougth better to satisfy the board.
From: atl2@lehigh.edu Newsgroups: comp.sys.next.programmer Subject: WOFrames incompatible with Netscape?? Help! Date: 6 Sep 1997 01:45:10 GMT Distribution: world Message-ID: <5uqcj6$255s@fidoii.cc.Lehigh.EDU> Hi all! I have entered a crucial stage in the development of a WebObjects application, only to discover that the WOFrames in my components fail to display properly on Netscape browsers. The same holds, incidentally, for the EmployeeBook example shipped with WO 3.1. All of these pages display just fine on OmniWeb. My server is a PPro running OS 4.1/mach with the apache httpd v. 2.0. Anyone know of a fix? God help me, I have a presentation coming up next week for which I MUST use Netscape... Cheers! Alex -- Alexander Levine Philosophy Department Lehigh University ATL2@lehigh.edu
From: scottm@nic.com (Scott) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 23:05:07 -0400 Organization: PETA (People for the Eating of Tasty Animals) Message-ID: <scottm-ya02408000R0509972305070001@news.erols.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit In article <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com>, LTaylor7@ix.netcom.com (Daniel L. Taylor) wrote: >Can't somebody sue him? Apple customers? Stockholders? Motorola? Can't >somebody drive him out before he kills Apple, the Mac, and Rhapsody? > If enough stockholders got together I imagine they could drive Jobs out (again). Look at it this way though, maybe Microsoft will buy Rhapsody and finally put out a good OS. :-/ -- -------------------------------------------------------------------------- Scott Maxwell - scottm (at) nic (dot) com Finger me for my public pgp key or http://www.nic.com/~scottm/key.html Remember when SPAM was the world's favorite mystery meat? Surfing is fast. The Internet is not. Therefore I "Sloth The Net"
Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc From: nagle@netcom.com (John Nagle) Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Message-ID: <nagleEG2JuC.LnM@netcom.com> Organization: Netcom On-Line Services References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <5upmg0$oun$1@halcyon.com> Date: Sat, 6 Sep 1997 04:15:48 GMT Sender: nagle@netcom6.netcom.com tzs@halcyon.com (Tim Smith) writes: >Lawson English <english@primenet.com> wrote: >>JObs was told by Motorola to stop playing games with clone-makers or face >>losing PowerPC's from Motorola. Jobs replied that he didn't care since Net >That is extremely unlikely. Almost all large companies that are involved >in more than one line of business let the separate business units make >mostly independent decisions. Unless the Motorola group that makes the >PPC chip is the same as the group that makes Mac clones (is it?), the >chip group probably couldn't care less if Apple screws the clone group, >as long as the chip group sells a lot of chips. Motorola has the problem that nobody else is buying the high-end PPCs in volume. They're becoming resigned to making microcontrollers based on various CPU cores; there's a 68HC11 line, a 68000 line, and and a PPC line of microcontrollers. But the markups on those things are nothing like the markups Intel gets on Pentiums, or Motorola gets on the high-end PPC parts. So there's not enough revenue coming in to keep up with Intel in that market. I wouldn't be suprised if they stopped further development of high-end PPC parts; who can they sell them to? On the other hand, if Apple sells x86 machines, they have a choice of making things that are basically Windows-compatible PCs, and having their software run on commodity PCs, or making an incompatible x86 machine, a strategy tried unsuccessfully by Texas Instruments, Wang, Sun, Radio Shack, and others. John Nagle
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 23:46:48 -0700 Organization: Netcom Message-ID: <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> In article <5upjgt$sme$1@news.lth.se>, d89cb@efd.lth.se (Christian Brunschen) wrote: > >Get ready for Steve's next big thing...a slotless, fanless, underpowered > >and overpriced computer with a little RAM, and even smaller HD (if one at > >all), and a permanent connection to a network where you wait for all your > >applications and data. > > I wonder where this rumor comes from. Where has Steve Jobs said anything > about stripping machines of powerful CPU, Memory, secondary storage ? > > Ah, I get it -- the letters 'NC' scare you. Well, a 'NC' need not be > slow, RAMless, or diskless. I have yet to see a NC spec which called for much more. You tried to refute me in your post by straddling the line between NC and PC, missing the point that PC will win out easily no matter how closely you straddle that line. > >Imagine! Everything that's yours stored somewhere else! No privacy! Web > >like wait times for your applications and files! A dead computer when the > >network goes down! > > Oh Please. Lots of people live quite comfortably with having their home > directories mounted over a LAN, using, say, NFS (under Unix). If you start > putting your servers and clients far apart, connected through 28k8 modems, > of course performancs will suffer. But put them on a 100Mb/s Ethernet, and > I doubt anyone will have any complaints -- especially since people will be > able to choose any one of a bunch of identical computers and get their own > personal computing environment. Depends on who you are and what you're doing. Believe me, plenty of people would complain and loudly. And what advantage will any NC offer here over a PC? None. > >It's the future people. A graphic designer won't mind waiting while > >his/her NC swaps MB after MB of program and file data back and forth over > >an Ethernet line to apply a single filter because MIS will save so much > >money on support costs. They will love NC's and jump right on them to save > >in support costs, just like they jumped right on Macintosh to save in > >support costs, right??? > > It's about adminitering all machines centrally. If all machines are NC:s - > in the sense that they aren't 'individual' machines, but rather 'clones' > regarding OS, software, and even most hardware - then you can shave off all > support costs regarding support of the 'individual' machines, since all > machines will be interchangable. Plus, you will be able to have a few 'fresh' > machines in 'reserve', so if any machine in the 'field' breaks down, you can > swap it for one that works, without further interruption to the work of > whoever is using the machine in question. And what advantage will NC's offer here over PC's? Are you seriously suggesting that PC's cannot be configured in the same way, especially as networking tools become more advanced? Again, how is Apple going to succeed here? > >They will gladly dump all their hardware just to be able to run slow Java > >apps on a slow NC...right? Or dump native apps in favor of Java > >interpreted apps served over a bogged down Ethernet line, right? Whoops! > >Some repair guy just smashed through a wall and the networking cable. > >Everybody go home now.... > > Um, Rhapsody runs, first and foremost, compiled Objective-C applications. > JIT-compiled Java is hardly 'interpreted' either, btw. It's also hardly as fast. > >And consumers will flock to this new technology. Imagine loading Word at > >28.8K. Heck, why doesn't AOL jump on this? All they have to do is > >integrate a word processor, spreadsheet, and DB into the AOL client > >software. Don't all of you want to do your work via AOL's server and store > >your data there? If that picture doesn't send chills up and down your > >spine, send your resume to Apple. They need people like you with "vision". > > You have heard the expression 'FUD', perhaps? Stands for 'Fear, Uncertainty, > and Doubt'. This is usually backed up with baseless assertions, sometimes > even outright lies. > > Just like you have been writing. Show me a spec which defines a full powered NC that is simply administered from a server...and I'll show you a Windows PC "administered" by Bill Gates. Steve has seen his vision...and he's going to watch as Gates passes it right over. -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 06 Sep 1997 00:02:35 -0700 Organization: Netcom Message-ID: <LTaylor7-0609970002350001@vic-ca1-24.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <5up1dc$ndg$1@news.lth.se> <341074F1.8F26E657@spamtoNull.com> In article <341074F1.8F26E657@spamtoNull.com>, Jonathan Harker <Limeyville@spamtoNull.com> wrote: > How long has NeXT for Intel existed as a marketable product? How well > has it done? There is no reason whatever to think that a Rhapsody for > Intel, devoid of the PPC and traditional Apple identity, will do any > better than NeXT, which had neither of those things, either. Oh, but wait! This time they are going to ship NextStep Intel on *NC's* instead of *PC's*...now it all makes sense, right? -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 05 Sep 1997 23:59:16 -0700 Organization: Netcom Message-ID: <LTaylor7-0509972359160001@vic-ca1-24.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <don_arb-0509971214210001@sea-ts1-p59.wolfenet.com> In article <don_arb-0509971214210001@sea-ts1-p59.wolfenet.com>, don_arb@wolfenet.com (Don Arbow) wrote: > Most of the hype about NC's deals with corporations and large intranets > where employees are using $2500 PC workstations to play solitaire. The > main push behind using NC's is that these companies can get rid of those > pricy machines, if the users are only going to run a word processor and a > database client. Centralized storage for programs and data makes sense > for maintenance and administration costs. > > No one has suggested that NCs will replace the "personal" market. > > As for storing of personal data on a remote server, quite a few people do > this very thing every day. I have 5 megabytes of storage available on my > ISP's computer, no one but myself can access it. > > Beware of the FUD... I love this...one person denounces my "FUD" by saying that NC's will be "...just as powerful..." as PC's, ignoring the point that if all you want is centralized administration that can already be done, so why go with an NC when you can go with a standard PC and have access to all that has already been built for it? Then you denounce my "FUD" by saying that NC's will be much cheaper...which means something has to go and the NC will not be just as powerful, making it a potentially useful product in some cases, but hardly the majority or all. If you are correct, then my post stands, not as FUD but as an accurate description for Apple's current customer base because to *us* that's how NC's look...useless, slow, bloated terminals. But the key thing here, you guys can't decide. The NC vision is a bit hazy and you guys need to figure it out better before pushing it as a solution.... Regardless of how NC's fly, I doubt Apple will be the provider. Do any corporations trust Apple at this point? Will they take Apple's solution or Microsoft's? Let's see, which will Motorola choose? PowerCC? IBM? Yale? Hmmm.... What, you thought Bill wouldn't have an answer for all of this? One things for sure...if this is indeed Jobs vision, it will be interesting to watch him go down as Billy Boy sweeps it up. -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 06 Sep 1997 00:04:58 -0700 Organization: Netcom Message-ID: <LTaylor7-0609970004590001@vic-ca1-24.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upkik$45m@dfw-ixnews10.ix.netcom.com> <5upv9c$6f5@spool.cs.wisc.edu> In article <5upv9c$6f5@spool.cs.wisc.edu>, schoinas@fox.cs.wisc.edu (Ioannis Schoinas) wrote: > Actually, when I can afford it, this is what I have in mind as a hobby > project. The technology is already available out there. Use old notebooks > as java terminals, a fast ethernet around the house, a wireless network > for things you want to carry with you, a big server, a cable modem to > reach out (or whatever comes out of the telcos), a computer with all > the 3d hardware with the TV. And consumers thought that PC's were complex.... -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 6 Sep 1997 20:14:31 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5urae7$f7c$1@attila.apana.org.au> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <5up1dc$ndg$1@news.lth.se> <341074F1.8F26E657@spamtoNull.com> In article <341074F1.8F26E657@spamtoNull.com>, Jonathan Harker <Limeyville@spamtoNull.com> wrote: :Christian Brunschen wrote: : :> :> Beginning of 1999 -- that's a little over a year away. By that time, :> Rhapsody Unified will have been shipping for several months, and :> Rhapsody Premier for almost a year. Most new application development :> for the Mac platform, should be happening under Rhapsody -- which :> means that, for new applications, indeed we do not care if we are :> running on PowerPC or on Intel. :> :> :> Rhapsody's cross-platform abilities mean that Apple is not 'locked' :> into the PowerPC processor anymore, and can act freely in regard to :> even someone like Motorola, their chip supplier for PowerPC. 'Freely' :> meaning, of course, 'in Apple's self-interest', as defined by Apple's :> leadership. : : :Yea, this sounds so nifty in theory. It *is* a fine concept, no doubt :about it. The central problem here, the fatal flaw, though, is that it :*completely ignores the character and the identity of the traditional :Apple market.* : :The "core identity" of the Apple market is that it *exists* as an :alternative to the Wintel market. It has always been so. Nine out of :every ten Apple customers will tell you this is why they use, and are :loyal to, Apple Computer. Many of them fully believe that the PPC :architecture is superior to Intel architecture, and so they make a :conscious choice to buy Mac, instead. Why are the alternative CPU choices limited to x86? OK, so Rhapsody is only being done for PPC and x86 at the moment, but in principle, it could be ported to almost anything - Apple machines with Alpha CPUs, anybody? I don't really beleive that Motorola would cut Apple off, though. Apple is the biggest customer for PPC chips (or at least, that's what I understand to be true) - lose Apple, and they lose PPC. Indeed, they completely give in to Intel. When push comes to shove, Motorola are going to have to keep supplying PPC parts to Apple for as long as Motorola makes them and Apple buys them. This is what Rhapsody buys Apple - independence from Motorola. The Wintel clone makers have had this for years now (able to buy from Intel, Cyrix, or AMD). There's no real reason why they have to stop with x86, either - Rhapsody could potentially support quite a lot of different CPUs, giving Apple (and cloners, if they're able to continue) lots of options. Sounds pretty good to me. -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: see-below@not-my-address.com (Matthew Vaughan) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 06 Sep 1997 03:48:03 -0700 Organization: MindSpring Enterprises Message-ID: <see-below-0609970348030001@ip183.mountain-view2.ca.pub-ip.psi.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <jrwst30-0509971132340001@jrwst30.j.resnet.pitt.edu> In article <jrwst30-0509971132340001@jrwst30.j.resnet.pitt.edu>, jrwst30@pitt.edu (James Wright) wrote: > i don't think that this is true. First of all, why would motorola do > this?If they were worried about the falloff from cloners being missed(they > only amount to about 300,000 machines), i think apple jumping ship would > put a little more of a dent in their pockets since they are going to ship > about 3 million machines this year. It looks like Apple may jump ship anyway. And already demand for desktop PPC CPUs is less than expected (read: the market is shirinking), and Motorola is converting some PPC capacity to other uses. It's very expensive to develop such chips, and those expenses can only be justified with a growing market. > secondly, IBM relies on the Power > Arhictecture for their workstation and supercomputing machines ; the next > generation of these machines will contain massively parralel > configurations of top of the line PowerPc's. But they already make seperate chips for their workstations and servers. They don't necessarily use the same PPC chips as PowerMacs. > It just doesn't > compute..lock this up in the rumor file. Lastly, if this is true, apple > may as well stop producing machines now. Who in the hwll would want a > machine that no one is going to produce software for. If apple > discontinues powerpc, and switches to intel machines, which platfor do > youthink the software is going to be written for?i know rhapsody is cross > platform, but the apps aren't; they still must be recompiled for both > intel and power pc. If this ever happened, apple would alienate alot of > customers who just purcased powerpc macs, and would lose them to Microsoft > simply because of stability. Apple won't stop making Macs because they want to fleece the sheep for as long as possible. While you, and many others, might rightly choose not to buy a Mac now, lots of others are more gullible, or still have pressing enough needs for a Mac to justify their purchase. But Apple needs the revenue (and hopefully profit) in the meantime. -- Matthew Vaughan matthewv at mindspring dot com (damn spammers...) --------- This is your mind. This is duality. This is your mind on duality...
From: see-below@not-my-address.com (Matthew Vaughan) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 06 Sep 1997 03:55:20 -0700 Organization: MindSpring Enterprises Message-ID: <see-below-0609970355210001@ip183.mountain-view2.ca.pub-ip.psi.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <LTaylor7-0509970612390001@vic-ca1-04.ix.netcom.com> <34100896.6928@A1.tch.harvard.edu> In article <34100896.6928@A1.tch.harvard.edu>, "Sean C." <Corliss@A1.tch.harvard.edu> wrote: > I'd like to see a source for this rumor. I sincerely can't believe that > they would do this. Why not? It's certainly a big gamble, but it actually makes a lot of sense. Of course, it generally means shafting all their existing Mac customers in the process... -- Matthew Vaughan matthewv at mindspring dot com (damn spammers...) --------- This is your mind. This is duality. This is your mind on duality...
From: wjtaggart@fuse.net (William J Taggart) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 06 Sep 1997 07:48:59 -0400 Organization: midwest payment systems Message-ID: <wjtaggart-0609970748590001@crazyhorse-19.fuse.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <jrwst30-0509971132340001@jrwst30.j.resnet.pitt.edu> <see-below-0609970348030001@ip183.mountain-view2.ca.pub-ip.psi.net> In article <see-below-0609970348030001@ip183.mountain-view2.ca.pub-ip.psi.net>, see-below@not-my-address.com (Matthew Vaughan) wrote: > In article <jrwst30-0509971132340001@jrwst30.j.resnet.pitt.edu>, > jrwst30@pitt.edu (James Wright) wrote: > > > secondly, IBM relies on the Power > > Arhictecture for their workstation and supercomputing machines ; the next > > generation of these machines will contain massively parralel > > configurations of top of the line PowerPc's. > > But they already make seperate chips for their workstations and servers. > They don't necessarily use the same PPC chips as PowerMacs. > I just finished an article about the new-and-improved IBM RS/6000. The article stated that the RS/6000 used 604e microprocessors (not clear about whether the 604e usage was part of the "new-and-improved"). How are these different from the 604e cpus used in PPCs? -- Bill Taggart <wjtaggart@fuse.net> Welcome to the Monkey House -- Kurt Vonnegut, Jr.
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <5urnfr$35a$7707@news.Kijfhoek.NL.net> Control: cancel <5urnfr$35a$7707@news.Kijfhoek.NL.net> Date: 06 Sep 1997 14:00:33 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5urnfr$35a$7707@news.Kijfhoek.NL.net> Sender: Funstore<Funstore@sod.nl> Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: hattonr@aug.com (Rick Hatton) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 6 Sep 1997 12:34:29 -0400 Organization: All USENET -- http://www.Supernews.com Message-ID: <19970906123429297977@ts2-06.aug.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <don_arb-0509971214210001@sea-ts1-p59.wolfenet.com> Don Arbow <don_arb@wolfenet.com> wrote: > Most of the hype about NC's deals with corporations and large intranets > where employees are using $2500 PC workstations to play solitaire. The > main push behind using NC's is that these companies can get rid of those > pricy machines, if the users are only going to run a word processor and a > database client. Centralized storage for programs and data makes sense > for maintenance and administration costs. We already have NCs. They are called dumb terminals. NCs will return us to the days when MIS directors controlled everything because it was all on the mainframe. If that isnt chilling to you, it is to me. half of my computer time is spent on a Vax or AS/400. Id much prefer a computer on my desk. .. and if someone is only playing solitaire.. not only is the computer not needed, the worker isnt either.
From: jurgen@oic.de (Juergen Moellenhoff) Newsgroups: comp.sys.next.programmer Subject: Re: pid of NSTask? Date: 6 Sep 1997 17:00:18 GMT Organization: OIC, Bochum, Germany Distribution: world Message-ID: <5us272$33n$1@nexus.oic.de> References: <3dgOoYuhHGqY@cc.usu.edu> In article <3dgOoYuhHGqY@cc.usu.edu> writes: > Is there a way to get the pid of a program > which was started using the launch method > of NSTask? I don't see a "pid" method in the > NSTask method list. Try the _undocumented_ method "- (int)_procid;". Bye, Juergen
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class (long) Date: 06 Sep 1997 19:58:01 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7u3fy2ul2.fsf@tom.ics.ele.tue.nl> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> <5umnvq$ju9$1@gte2.gte.net> <34100283.0@epflnews.epfl.ch> <5uph79$pvu$1@gte1.gte.net> In-reply-to: ga@ed4u.com's message of 5 Sep 1997 17:58:01 GMT In article <5uph79$pvu$1@gte1.gte.net> ga@ed4u.com (G. Apple) writes: Operator overloading and multiple inheritance are two of the most powerful features of C++, and I just hate like hell to lose them in the stampede to Java and Objective C. I `grew up' in the Objective-C camp of the OO world, and after 4 years of experience with the language, I developed TOM. TOM started out as a better Objective-C and it was only because I wanted extensions (the TOM word for Objective-C categories; they can add methods and instance variables) that I added multiple inheritance (MI) to the language: the necessary mechanisms were already present to support extensions, and adding MI was, sort of, trivial. MI in TOM isn't as baroque in Eiffel or painful as in C++; it is best described in C++ speak as `virtual public'. In the past 18 months I've written +/- 500 classes, some trivial, some not, and armed with this experience, I agree that multiple inheritance, at least in the way TOM provides it, is very useful. To speak of MI as one of the most powerful features (of C++) doesn't do right to truely the most powerful aspect of object oriented programming: polymorphism. It is in this field that C++ fails miserably. Sure, you can get polymorphism if you need it, but only if you own the code of every superclass so you can add `virtual' to those methods needing it. And polymorphism is only one example of the fact that C++ provides many things that you `can get' but don't _have_. --Tiggr
From: marssaxman@sprynet.com.antispam (Mars Saxman) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 06 Sep 1997 11:17:31 -0800 Organization: Red Planet Software Message-ID: <marssaxman-0609971117310001@hd55-082.hil.compuserve.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <acurylo-0509970959590001@p2-29.van.tvs.net> In article <acurylo-0509970959590001@p2-29.van.tvs.net>, acurylo@inmediapresents.nospam.com (Alex Curylo) wrote: > In article <B03540AE-CE02A@206.165.44.43>, "Lawson English" > <english@primenet.com> wrote: > > > JObs was told by Motorola to stop playing games with clone-makers or face > > losing PowerPC's from Motorola. Jobs replied that he didn't care since Net > > Computers could just as easily run on x86 CPUs. > > Well, well, well, well, well. It's nice to hear even a wild rumour that > backs up exactly what I've been saying on the Knife boards for a while > now. > > Apparently S. Jobs shares my conviction that PPC hw is going to have all > the market appeal in six months of a NeXTStation post-NeXTStep/Intel... > > ---- > Alex Curylo -- alex@witty.com The really pathetic thing about this point of view is quite apparent to anyone familiar with the assembly language on both the PowerPC and the Pentium. PowerPCs are a fairly elegant, orthogonal, compiler-friendly, and generally very nice and advanced architecture. Pentiums are an archaic conglomeration of ancient kludges stacked one atop another with no regard whatever for programmer hair loss. Yes, the PPC has its flaws; but it is difficult to overstate the level of ugliness presented by the x86 command set. It's a pity there aren't more Mac compiler authors; otherwise we could threaten to lynch anyone who suggested abandoning the PowerPC. -Mars
From: Jonathan Harker <Limeyville@spamtoNull.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: Sat, 06 Sep 1997 15:32:40 -0400 Organization: Netcom Message-ID: <3411AFD8.8EF55F7D@spamtoNull.com> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> <mschemen-0409972046570001@ts005d05.ftw-tx.concentric.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Mathew Schemenaur wrote: > > > > > "To buttress their argument, they noted that the licensees have been paying > > only about $50 per system for their Mac OS licenses. They neglected to > > mention the separate hardware license fees the cloners have also been > > paying -- in some cases, more than $180 per unit. (Note that most of this > > money goes straight to Apple's bottom line, since there's no cost of goods > > nor any new R&D expenditure required.) > > If this is true then every sell of the MacOS goes straight to Apple's > bottom line. Forget about the huge investment in the original designs, the > PowerPC conversion, PCI switch over, ... And if what you say is true, then every person who bought 0S8 off the shelf was likewise subsidized by Apple to the tune of "several hundred dollars" since, like clone companies, they bought no new Apple hardware to go along with it. Secondly, your viewpoint is incorrect because it assumes Apple received no payment for the OS from clone companies. In fact, not only did the clone companies pay $50 for OS 7.6, they also paid an average of $150 per system to Apple for hardware licensing fees. That means, according to you, that Apple "subsidized" off-the-shelf OS8 buyers for even *more money* than it "subsidized" clone companies. Of course, this is all balderdash because Apple subsidized nothing to anyone. Apple was paid by everyone, not the other way around. Is it your suggestion that, in the absence of cloners, Apple will now SLASH R&D expenditures to eliminate the "subsidies" it had been "paying" to cloners? Let's see, they ought to be able to cut back R&D expenses by about half now that the cloners are gone. Yea, right. > > Of course they are going to compare it to the legal contract and not an > agreement that was never finalized by a CEO that is no longer with Apple. > That does not make it a lie. No, but it does make it a clear breach of contract. The pre-agreements were written, signed documents that were *binding* on not just the cloners, but on Apple as well. Why do you think Apple paid off Power to the tune of a whopping $100M? Apple may not be so lucky with Motorola. > > According to the expert Harry Norr, Apple was already getting 100 to 280 > per box when you concider software and hardware fees. So a jump to 100 to > 500 does not seem like a jump at all. Right. Sure. Sell one million machines @ $280 and pay Apple $280M. Or pay $500M on the same number of sales. A paltry difference of only 220 MILLION DOLLARS. Yea, buddy--no jump at all. > > I think Gil got fired because he agreed to terms that were not much better > then the original terms. At least not enougth better to satisfy the board. Don't forget the "board" got "fired" too--or most of them. -- Jonathan Hollingsworth Harker New astronomical data shows the age of the universe as measured by the latest galactic red-shift data at between 8 and 12 billion years. However, hydrogen decay data proves indisputably that many stars are 15-20 billion years of age. Modern astronomical theory about the formation and age of the universe is so accurate that it tells us that many stars are older than the universe in which they exist. And you laughed when they said the world was flat?
From: ga@ed4u.com (G. Apple) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class (long) Date: 6 Sep 1997 19:51:35 GMT Organization: Advanced Communications Engineering, Inc. Message-ID: <5usc87$2fq$1@gte2.gte.net> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> <5umnvq$ju9$1@gte2.gte.net> <34100283.0@epflnews.epfl.ch> <5uph79$pvu$1@gte1.gte.net> <x7u3fy2ul2.fsf@tom.ics.ele.tue.nl> In article <x7u3fy2ul2.fsf@tom.ics.ele.tue.nl>, tiggr@ics.ele.tue.nl (Pieter Schoenmakers) wrote: > I `grew up' in the Objective-C camp of the OO world, and after 4 years of > experience with the language, I developed TOM. TOM started out as a > better Objective-C and it was only because I wanted extensions (the TOM > word for Objective-C categories; they can add methods and instance > variables) that I added multiple inheritance (MI) to the language: the > necessary mechanisms were already present to support extensions, and > adding MI was, sort of, trivial. MI in TOM isn't as baroque in Eiffel or > painful as in C++; it is best described in C++ speak as `virtual public'. > In the past 18 months I've written +/- 500 classes, some trivial, some > not, and armed with this experience, I agree that multiple inheritance, > at least in the way TOM provides it, is very useful. > > To speak of MI as one of the most powerful features (of C++) doesn't do > right to truely the most powerful aspect of object oriented programming: > polymorphism. It is in this field that C++ fails miserably. Sure, you > can get polymorphism if you need it, but only if you own the code of > every superclass so you can add `virtual' to those methods needing it. > And polymorphism is only one example of the fact that C++ provides many > things that you `can get' but don't _have_. --Tiggr Coming from use of MacApp all the way from it's Pascal days until present, I haven't run into this problem much, although I certainly acknowledge it's existence. Everything in Object Pascal was public and overridable. Although some effort has now been made to privitize (and protect) some things in the present C++ version, most of MacApp's methods are still virtual and accessible. One thing I noticed long ago with MacApp's (original) single inheritance was that subclasses were often introduced simply to add orthogonal functionality that is now much better implemented through multiple inheritance. Plus, the MI-ed classes are then available for use elsewhere without the inherited baggage they would have otherwise. Properly used MI actually results in much better code organization and readibility because of the ability to encapsulate individual functionality without unnecessarily entangling it with unrelated issues. The code becomes much more modular and reusable at a lower level. Having started with SI and then moved to MI (and having seen and used many good examples having proper use), it's incomprehensible to me to observe particularly the Java crowd disparaging MI so vehemently. Often, MI does provide a way to extend someone else's library through overrides and MI extension -- but yes, it does often require that they cooperate by making the necessary methods virtual. -- G. Gordon Apple, PhD The Ed4U Project Advanced Communications Engineering, Inc. Redondo Beach, CA ga@ed4u.com
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 6 Sep 1997 15:32:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B0372D47-F4AD@206.165.43.161> References: <5urae7$f7c$1@attila.apana.org.au> nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.system, nntp://news.primenet.com/comp.sys.mac.misc MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit > I don't really beleive that Motorola would cut Apple off, though. Apple is > the biggest customer for PPC chips (or at least, that's what I understand > to > be true) - lose Apple, and they lose PPC. Indeed, they completely give in > to Intel. When push comes to shove, Motorola are going to have to keep > supplying PPC parts to Apple for as long as Motorola makes them and Apple > buys them. Nope. Ford Motor, and various other embedded market companies provide far greater markets for PPC's than APple's measely little 4 million/year. And the need to make things faster, faster, isn't quite as frantic as it is on the desktop CPU side. IBM has a need to keep the workstation CPUs going faster, since they use the PowerPC in their own workstations, but IBM-only gives Apple a single source for the CPU, making them more expensive. ------------------------------------------------------------------------- Build a better mousetrap, yes, but then market it like crazy, or instead of catching mice you'll only collect dust. -David Yeargin -------------------------------------------------------------------------
From: gtupar@ctp.com (Georg Tuparev) Newsgroups: comp.sys.next.programmer Subject: Re: Objective-C Fragile Base Class Date: 6 Sep 1997 23:09:31 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5usnrb$6lv$1@concorde.ctp.com> References: <340f1650.0@cisun2000.unil.ch> ga@ed4u.com (G. Apple) wrote: > In article <340e6ca4.0@epflnews.epfl.ch>, marco@sente.ch (Marco Scheurer) wrote: > > Oh yes? I have never seen a geometry text using + like that. > > > It's obvious that you've never spent time > working with vectors, vector spaces, or abstract algebra. The concept of > an "operator" is something that is well understood by people versed in > mathematics, which often unfortunately does not include programmers. If this matter, I have a master in math, and can tell you that you have no idea what you are talking about. And from Marco you can only learn! Just my $0.02 -- /\/\ Georg Tuparev <georg_tuparev@ctp.com> | Currently in Dublin / /_ \ Cambridge Technology Partners 118/119 Lower Baggot Street \ / / Apollo House, Apollolaan 15 Dublin 2, Republic of Ireland \/\/ 1077 AB Amsterdam, The Netherlands Tel: +353(1)607-9083 Tel: +31(20)575-0492 Fax: +31(20)575-0500 WWW: http://www.ctp.com Hiroshima 45, Tschernobyl 86, Windows 95, Office 97
From: lozinski@openstepnews.com (Christopher Lozinski) Newsgroups: comp.sys.next.advocacy,comp.sys.next.marketplace,comp.sys.next.programmer,comp.sys.next.sysadmin,comp.mac.advocacy Subject: OpenStep Newsletter Date: 6 Sep 1997 23:53:56 GMT Message-ID: <5usqek$ft5$1@nntp2.ba.best.com> 1. Connecting One Mach Computer to the Internet 2. Connecting a Network of Mach Computers to the Internet 3. Uploading Email to the ISP 4. Downloading Email from the ISP 5. 213 OpenStep/WebObjects Developers Available 6. Apple Buys Power Computing 7. Steve Jobs is Brilliant 8. What is Going on at Apple? 9. Sun Dropping OpenStep 10. OpenStep Jobs a)Great for "People Persons" b)Object Director c)Virginia d)San Francisco Bay Area e)Go Anywhere Consultants 11.NeXT Monitor on P.C. 12. BPG News 13. Mitch Manditch Promoted 14. Faxing From a Portable 15. Big Green CD 7.0 16.Common Hardware Reference Platform (CHRP) 17. Project Design Reviews 18.Typing Injury FAQ 19. Subscriptions Now Available 1. Connecting One Mach Computer to the Internet With the expected arrival of Rhapsody, new life has been breathed into the Mach Operating System. Rather than migrating off of Mach, people are investing in their Mach environments. One important area of development is a method for connecting Mach based computers to the Internet. Here is the basic approach. For a single Mach-based computer it is reasonably easy to connect to the Internet over a phone line using PPP. Once that is done, all of your internet applications should work immediately except NeXTMail. The simplest way to send and receive email across PPP is to use the Eloquent browser. <http://www.take3.com/> Eloquent makes it easy to send outbound messages with the smtp protocol, and to receive inbound messages from a POP server. Eloquent also supports the IMAP protocol. Unlike NeXTMail, Eloquent talks SMTP to the ISP's outbound mail server and POP3 to the ISP's inbound mail server, eliminating the need to configure Sendmail and figure out how to download email messages. 2. Connecting a Network of Mach Computers to the Internet There are several ways to connect a Mach network to the Internet. If you have money and are willing to spend a few hundred dollars per month, just connect your Ethernet to an ISDN line or to a frame relay through a bridge or router. Your Internet Service Provider (ISP) will give you IP addresses for all of your computers, you will configure the local routing, and all of your Internet applications will work. On the other hand, if your budget is smaller, or you live in Europe where connectivity costs are higher, then try the following approach. It is possible to connect a network of computers across a single phone line to an ISP. If you are running PPP from a gateway computer over a single telephone line, you need to route all traffic through the gateway computer. There are two ways to do this, you can use static IP addresses or dynamic IP addresses. With static addresses, it is possible to have your ISP give you several ip addresses, one for each computer, and then you can configure your network to use them. You will be fully on the Internet, and can use all of the standard software painlessly. Alas this option is expensive, about $495 setup plus $145/month from my ISP. That price is a lot more than the standard $25/month dynamic IP Internet connection. Such discount connections cannot easily be used for a network because they are typically a single dynamically assigned Ip address. The ISP's computers do not know about all of your computers; they think that just a single computer is being served. Although your network computers will be able to route messages through the gateway, and send messages out to the Internet, messages inbound will not find the desired host because the IPS's computers only know of one host. The best way to run a network across an inexpensive phone connection is to have the gateway computer run proxy applications for the other computers. Thus, the email server runs on the gateway computer and forwards and receives email to the ISP. A proxy http server runs on that computer. For every Internet application, there needs to be a proxy process on the gateway computer that forwards and returns messages appropriately. 3. Uploading Email to the ISP One Obvious gateway computer application is Sendmail. It comes with the Mach OS and is important because it forwards and dispatches email. An updated version is available at ftp://ftp.peak.org/pub/next/apps/mail/. Sendmail is historically at the heart of Internet mail. It was initially developed by a college student as a way to deal with the rapidly evolving nature of Internet email addresses. At the time, there were many different email standards and the rules were changing daily. This student was studying expert systems and built a complete expert system to figure out how to dispatch email messages. As the rules changed, his software kept up to date with the needs. Of course today sendmail is no longer an appropriate tool for small networks. Despite major changes to the user interface, it is just too complicated for the average small business. Email addresses have standardized, the ISP can deal with any complexities. Security is a huge issue when connecting to the Internet, and Sendmail has numerous loopholes. Performance is also an issue when running mailing lists, and again, Sendmail has lower performance than competing packages. Alternative choices for an email agent are Smail and Qmail <http:/www.qmail.org>. I could not find either one publicly available and precompiled for Mach and tested, but I have found individuals who are using each of the products on Mach. Reportely Qmail has no NetInfo support, and is not plug compatible with sendmail, so I recommend smail for this application. Would anyone be willing to create an installation package for one of these tools, and post it on the internet? Let me know what you thought about this article. Its style is different than what I have traditionally done. Would you like to see more articles like this? 4. Downloading Email from the ISP SendMail and Smail solve the problem of sending mail from a network of computers, or from a single computer running NeXTMail, but they do not help with the problem of downloading email. There are three solutions to that problem. By now few people are running UUCP, so the original NeXT tools are not much help. Popover reaches out to a POP server and downloads mail into the file system for access by the local computers. It is simple and easy to use in Mach. For a more advanced solution try Fetchmail. Fetchmail runs on all the Unix machines, including Mach, but does not have a Mach GUI. Fetchmail talks any variant of POP or IMAP to the server, and forwards the downloaded mail to the local SMTP mail listener. You can find the Fetchmail home page here: <http://locke.ccil.org/~esr/fetchmail/>. You can download Fetchmail for mach here: <ftp://ftp.peak.org/pub/next/apps/mail/>. 5. 213 OpenStep/WebObjects Developers Available We now have 213 OpenStep/WebObjects developers, system administrators, QA people, managers, and architects registered in our database. I have never before managed to maintain so many active candidates properly registered in the database. When companies call in now to hire someone, we can provide them with numerous excellent candidates on a same day basis. If you need OpenStep/WebObjects developers let me know. <lozinski@openstepnews.com>. You will be amazed by the calibre of the individuals I know, and by the quality of the service provided. 6. Apple Buys Power Computing Apple has acquired Power Computing, the largest clone vendor, for $100 million in apple stock. Power Computing will continue to ship Motorola-based hardware until the end of this year. After that it will ship Intel computers, presumably with the Rhapsody Operating System. Power Computing's first WinTel notebooks will be announced on Monday, and will begin shipping this month. 7. Steve Jobs is Brilliant Whereas most are criticizing Apple for its current decisions, I think I see the strategy at play here. Let me lay it out and, with your feedback, we will see if my crystal ball is any good. I believe that Apple plans to use its proprietary technology as a competitive advantage in selling Intel-based and Motorola-based hardware. First Steve negotiates a deal with Microsoft, ensuring that the critical applications of Excel and Word will continue to be available on the Macintosh Operating System. That solidifies Apple's position, so that it is not as dependent on the other third-party software vendors. That allows Apple the freedom to flex its muscles in the market. Whereas the previous MacOS can be purchased with hardware from the clone vendors, MacOS 8 can only be purchased bundled with Apple (and now Power Computing) hardware. To buy a clone with Mac OS bundled ($50 to apple) and add an aftermarket Mac OS 8 ($95) and spend an hour doing it ($100) gives Apple a $250 dollar competitive advantage against the clone makers while not closing them out of the market. I suspect that Apple will eventually license OS8 to the clone vendors, probably about the time that Rhapsody ships. On the Intel side, Power Computing will start manufacturing Intel-based computers beginning next year, at about the time Rhapsody is available. I expect that Apple/Power Computing will be the only company bundling Rhapsody on Intel hardware. Again, I have no access to the Apple Board Room, so this is purely my speculation based on the news available to all. Your comments appreciated. 8. What is Going on at Apple? In general I have chosen a policy of not covering the activities at Apple, except as they pertain to Rhapsody. But the following item is just too odd to be ignored. Steven Jobs, in an email to all of his employees, explained his reasons for purchasing Power Computing. It was a very Macintosh centric memo with no mention of Rhapsody. You can read it at http://www.sjmercury.com/business/apple/memo090297.htm The top of the email gives a confusing message: Subject: Apple Acquires Power Computing From: Steve Jobs, sjobs@apple.com To: apple_employees$@group.apple.com A P P L E C O N F I D E N T I A L P L E A S E D O N O T F O R W A R D A P P L E C O N F I D E N T I A L P L E A S E D O N O T F O R W A R D Did he want this message kept confidential? Surely Steve is smart enough to know that such an email would be forwarded to the press and given its due attention. It reminds me of his last such "confidential" email to the staff at Pixar, assuring them that he planned to stay at Pixar and did not plan to be the CEO of Apple. I wish I knew what this meant. 9. Sun Dropping OpenStep The following story is entirely pieced together from bits and pieces I have heard on different days. I am confident that there are specific errors in the details, but that the general story is correct. Sun's management is very focussed on Java, and the company is dropping OpenStep in favor of Java. First they changed the name of ObjectPlan to JavaPlan, and made it work with Java source code as well as with OpenStep applications. Then they ported JavaPlan to Java. Then Sun moved all 20 OpenStep developers to Lighthouse to work on JavaPlan (in Java). Then Sun absorbed Lighthouse into Sun. Meanwhile, rumor had it that Sun was trying to sell off the source code to its Lighthouse applications. I also heard a rumor that they were giving away license keys to their applications for free. There was also discussion on Usenet of what was happening to the joint Sun-NeXT OpenStep specification. My take is that it is no longer relevant. 10. OpenStep Jobs If you are interested in any of the following positions, please send a note to: <lozinski@openstepnews.com>. Even if you are already registered in the database, please let me know of your interest in a particular spot, or even if you are generally interested in what positions may come up. It is only when I hear from you that I know that you are interested. a)Great for "People Persons" If you happen to be a OpenStep developer, but you think that your true calling is in dealing with people, then we have a job for you. This job requires you to help others, train them, write demo programs for them and let them know that their issues are being addressed. If you are able to tune into what someone else's issues are, and can address those issues efficiently, then you will love this position. The best thing is that there are no-long term deliverables; your managers will not measure you on deliverables. It is all about people, but you need to be a developer first and foremost. b)Object Director This is one of the more senior positions to cross my monitor. This company wants a senior manager to take control of an entire group of 30 developers and be in charge. The company needs someone who can manage such a large group and make things happen. The leverage is huge, and they are willing to pay what it takes to recruit such a person. Without saying where the position is located, I will say that it is a popular destination among the people who apply for positions. c)Virginia Virginia is a hot destination again this month, and it has nothing to do with the usual companies. If you like Virginia I have a choice of companies/environments where I can place you, based on what your preferences are. d)San Francisco Bay Area Last month a major NeXTStep company laid off much of its work force, creating a short-term glut of OpenStep developers. But now most have been hired, and demand for developers in the Bay Area continues to be strong. I have one client who is interviewing people this week, but they only want people who are currently living close to the Bay Area. e)Go Anywhere Consultants We have a steady demand for consultants who are willing to pack their bags and go to this month's hot destination. While new destinations may arrive on a one-day-notice basis, this month's hot destination is Los Angeles. 11.NeXT Monitor on P.C. If you have slow old black hardware, but a wonderful monitor, and you want to run the same monitor on white hardware, try using the following video card: <http://www.photonweb.com/>. 800-547-2349. 12. BPG News Last month I announced that I had hired a Human Resources Professional Vernetta Williams. Having her here has really helped my services. I simply could not handle the volume of resumes and jobs without her. With her, I am on top of my work on a daily basis. The volume of resumes correctly indexed in the database is way up. The number of developers that we know is up. The number of hiring companies that I am in touch with is up. BPG is well on its way to becoming the hottest market for OpenStep/WebObjects Developers. It has been great working with her. Furthermore she brings a Human Resource perspective that I never had to this job function which I never had. Certain aspects of the process are being measurably improved because of her input. 13. Mitch Manditch Promoted Apple has promoted Mitch Manditch to head of Sales in the Americas. This territory includes North, South, and Central America. Prior to the aquisition of NeXT, he was Vice President of Sales at NeXT. This promotion increases the presence of NeXT Inc. alumni at the senior levels in Apple. 14. Faxing From a Portable Tim Finnegan of WorkStation 2000 <w2000@cts.com> writes: When clients buy a portable, they frequently need fax capability running on the new PCMCIA fax modems. Alas, we have discovered that NXFax works with the 14.4Kbaud modems, but needs to be updated to work with the 28.8 modems. Furthermore, it is becoming increasingly difficult to purchase 14.4 PCMCIA modems. When questioned about this issue, the vendor of NXFAX sales@bandw.com wrote: At this point it is not likely that we will be adding new features or releasing new versions of NXFax. We do give it thought, but over the last couple of years we have had to move on to other things, making it difficult to move back. There is a real need for a fix to this problem. Since this article was first written Black and White Software has quite selling NXFAX, and has made it publicly available. 15. Big Green CD 7.0 The Big Green CD is a very nicely organized collection of NeXTStep-related materials. I was finally able to find the BackSpace Screen saver on their discs after looking in vain in all the usual places. They have the entire comp.sys.next.* Usenet hierachy indexed for searching with Digital Librarian. If you are interested in more informaton check out <http://www.skylee.com/> or send email to <disc@skylee.com> 16.Common Hardware Reference Platform (CHRP) The common hardware reference platform is a motherboard architecture designed to allow Apple-compatible computers to use Intel-compatible components. The volume and variety of Intel components are much greater than there are of apple components and the price of Intel components is much lower. With the CHRP architecture, it is possible to integrate those components into an Apple Macintosh, or later into a Rhapsody computer, providing greater value to the customer. Mac OS8 runs on CHRP hardware, and presumably Rhapsody will as well. A number of companies own CHRP designs, but none of them can now license Mac OS8, nor can they yet license Rhapsody, so I assume that only Apple/Power Computing will manufacture such designs. 17. Project Design Reviews If your project is scheduled for a design review, you may be wondering what questions to discuss. For some suggestions, check out: <http://members.home.net/slater/previews/index.htm> 18.Typing Injury FAQ I know far too many developers who have typing injuries, even shoulder injuries. If you have concerns, check out the Typing Injury FAQ at: Wright, K.S. & Wallach, D. S. (1997) "Typing Injury FAQ " http://www.cs.princeton.edu/~dwallach/tifaq/ 19. Subscriptions Now Available This newsletter is now available as ASCII email, NeXTMail, and MIME mail. Send your subscription and cancellation requests to the email address corresponding to the version that you want or do not want: ASCII version newsletter-request@lists.best.com NeXTMAIL version nextletter-request@lists.best.com MIME version newsletter-mime- request@lists.best.com The body of the message should contain the keyword 'subsingle' alone or, if you want to subscribe a third-party address, the subsingle keyword may be followed by an email address. subsingle or subsingle email-address You will then be subscribed. (Note: Do not use the 'subscribe' keyword, or you will get the digest version of the newsletter once a year). You can also subscribe multiple people in a single email message, one per line. Since this newsletter targets OpenStep Developers, I prefer if you subscribe to the NeXTMAIL version. If you are able to do that, then please tell me a little about how you are using NeXTStep or what you would like to see happen. To make my job easier, please left-justify your name and email address, one line each, as follows: Christopher Lozinski BPG 35032 Maidstone Court Newark, CA 94560 lozinski@OpenStepNews.com Copyright 1997 Berkeley Productivity Group. All rights reserved. Non-profit, non-commercial publications and web sites may reprint or link to articles if full credit is given. Others please contact us. We do not guarantee accuracy of articles. Caveat lector. Publication, product, and company names may be registered trademarks of their companies.
From: uli@tallowcross.uni-frankfurt.de (Uli Zappe) Newsgroups: comp.sys.next.programmer Subject: Services: "NXAsciiPboardType" <> "NeXT plain ascii pasteboard type" Date: 6 Sep 1997 23:52:48 GMT Organization: Frankfurt University Computing Center Message-ID: <5usqcg$ach@tallowcross.uni-frankfurt.de> Hi, I have a question concerning the Send Type and Receive Type for services. While NEXTSTEP 3.3 documentation refers to the types in the style of "NXAsciiPboardType" a lot of services actually declare types in the style of "NeXT plain ascii pasteboard type" which I couldn't find documented anywhere. Does anybody know where this alternative style stems from (NEXTSTEP < 3?), and, more important, has a list of all possible types in this style? Stangely enough, the performServices utility (which is found e.g. in the TickleServices distribution) seems to work *only* with this alternative style... Thanks for your help! Bye Uli -- _____________________________________________________________________ Uli Zappe E-Mail: uli@tallowcross.uni-frankfurt.de (NeXTMail,Mime,ASCII) PGP on request Lorscher Strasse 5 WWW: - D-60489 Frankfurt Fon: +49 (69) 9784 0007 Germany Fax: +49 (69) 9784 0042 staff member of NEXTTOYOU - the German NEXTSTEP/OPENSTEP magazine _____________________________________________________________________
From: scott@doubleu.com (Scott Hess) Newsgroups: comp.sys.next.programmer Subject: Re: Automatic spinning cursor of Mach not going away ?? Date: 7 Sep 1997 03:04:58 GMT Organization: Is a sign of weakness Message-ID: <SCOTT.97Sep6223450@slave.doubleu.com> References: <5uje9s$1q2$1@concorde.ctp.com> <34104142.10728997@news.wco.com> <5uq102$91q$1@concorde.ctp.com> In-reply-to: tsengel@immd9.informatik.uni-erlangen.de's message of 5 Sep 1997 22:27:14 GMT In article <5uq102$91q$1@concorde.ctp.com>, tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) writes: In order to get a good responsiveness out of TheShelf I was making my [NSWorkspace lauchThisDoThatTakeYourTime] request in a separate thread. Even if the AppKit is not thread save this works quite nicely... <...> So the current workaround is the old performSelector:withObject:afterDelay:. But this means NSWorkspace blocks :-( <...> P.S: Ok...there are other possible no-blocking solutions as well...but all of them are more work then I'm willing to invest right now. The easiest solution _I_ can see would be to put the NSWorkspace calls into a seperate process entirely, and let _it_ block. Just talk to it using oneway methods only. Sure, it's not lightweight, but if you're only running one instance, big deal... Later, -- scott hess <scott@doubleu.com> (606) 578-0412 http://www.doubleu.com/ <Favorite unused computer book title: The Demystified Idiots Guide to the Zen of Dummies in a Nutshell in Seven Days, Unleashed>
From: sanguish@digifix.com Newsgroups: comp.sys.next.advocacy,comp.sys.next.announce,comp.sys.next.hardware,comp.sys.next.software,comp.sys.next.misc,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.sys.next.programmer Subject: NEXTSTEP/OpenStep Resources on the Net Supersedes: <16621873000033@digifix.com> Date: 7 Sep 1997 03:52:35 GMT Organization: Digital Fix Development Message-ID: <8835873604833@digifix.com> Topics include: Major OpenStep/NEXTSTEP World Wide Web Sites OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups Major OpenStep/NEXTSTEP FTP sites NeXTanswers Major OpenStep/NEXTSTEP World Wide Web Sites ============================================ The following sites are a sample of the OpenStep related WWW sites available. A comprehensive list is available on Stepwise. Stepwise OpenStep/Rhapsody Information Server http://www.stepwise.com Stepwise has been serving the OpenStep/NEXTSTEP community since March 1993. Some of the many resources on the site include: OpenStep Third Party Software guide, Developer Directory, Mailing List information, extensive listing of FTP and WWW sites related to OpenStep and NEXTSTEP, OpenStep related Frequently Asked Questions. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. This is the World Wide Web interace to the FTP site. Apple Enterprise Software Group (formerly NeXT Computer, Inc.) http://www.next.com Here is where you'll find the NeXTanswers archive, with information on OpenStep installation, drivers and software patches. Apple Computer's 'Prelude to Rhapsody' Self Support Site http://devworld.apple.com/dev/prelude.html This site has been constructed to help you help yourself to learn as much as possible about the foundation for Rhapsody, today's OPENSTEP. The site provides an informal collection of pointers, references, and starting points for developers who are using the Prelude to Rhapsody bundle, distributed at this year's Worldwide Developer Conference. OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups ==================================================== COMP.SYS.NEXT.ADVOCACY This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. COMP.SYS.NEXT.ANNOUNCE Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. COMP.SYS.NEXT.BUGS A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT-specific groups as well. COMP.SYS.NEXT.HARDWARE Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. COMP.SYS.NEXT.MARKETPLACE NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. COMP.SYS.NEXT.MISC For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! COMP.SYS.NEXT.PROGRAMMER Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. COMP.SYS.NEXT.SOFTWARE This is a place to talk about [third party] software products that run on NEXTSTEP systems. COMP.SYS.NEXT.SYSADMIN Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. ** RELATED NEWSGROUPS ** COMP.SOFT-SYS.NEXTSTEP Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. COMP.LANG.OBJECTIVE-C Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. COMP.OBJECT Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep next-advocacy next-announce next-bugs next-hardware next-marketplace next-misc next-programmer next-software next-sysadmin object lang-objective-c (For a full description, send mail to listserv@antigone.com). The subscription syntax is essentially the same as Majordomo's. To subscribe, send a message to *-request@lists.best.com saying: subscribe where * is the name of the list e.g. next-programmer-request@lists.best.com Major OpenStep/NEXTSTEP FTP sites ================================= ftp://ftp.next.peak.org The main site for North American submissions formerly ftp.cs.orst.edu ftp://ftp.peanuts.org: (Peanuts) Located in Germany. Comprehensive archive site. Very well maintained. ftp://ftp.nluug.nl/pub/comp/next NeGeN/NiNe (NEXTSTEP Gebruikers Nederland/NeXTSTEP in the Netherlands) ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! _________________________________________________________________ Written by: Eric P. Scott ( eps@toaster.SFSU.EDU ) and Scott Anguish ( sanguish@digifix.com ) Additions from: Greg Anderson ( Greg_Anderson@afs.com ) Michael Pizolato ( alf@epix.net ) Dan Grillo ( dan_grillo@next.com )
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 7 Sep 1997 15:25:16 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5utdrs$41l$1@attila.apana.org.au> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> In article <B0372D47-F4AD@206.165.43.161>, Lawson English <english@primenet.com> wrote: : : : :> I don't really beleive that Motorola would cut Apple off, though. Apple :is :> the biggest customer for PPC chips (or at least, that's what I understand :> to :> be true) - lose Apple, and they lose PPC. Indeed, they completely give :in :> to Intel. When push comes to shove, Motorola are going to have to keep :> supplying PPC parts to Apple for as long as Motorola makes them and Apple :> buys them. : :Nope. Ford Motor, and various other embedded market companies provide far :greater markets for PPC's than APple's measely little 4 million/year. Thanks. I wasn't aware of this. :And the need to make things faster, faster, isn't quite as frantic as it is :on the desktop CPU side. : :IBM has a need to keep the workstation CPUs going faster, since they use :the PowerPC in their own workstations, but IBM-only gives Apple a single :source for the CPU, making them more expensive. Are the desktop PPC's losing money for Motorola? Are they making money? Even if they're only making a modest profit on them, they'd be crazy to stop. Still, with Rhapsody being reasonably cross-platform, Apple could get away with sourcing PPC parts from IBM only, as they'd have the option of producing systems with other CPUs (so IBM PPC parts would be competing with Intel, Cyrix, AMD, IBM x86, maybe even Sparc and Alpha - if the Linux community can cover those platforms, then surely a multi-billion-dollar corporation could manage it?). Maybe Apple just need to hire Linus Torvalds to run their OS-development. :-) Matt -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: 00093182@bigred.unl.edu (Josh Hesse) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 6 Sep 1997 08:06:55 GMT Organization: University of Nebraska--Lincoln Message-ID: <5ur2uv$r3s@crcnis3.unl.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <acurylo-0509971010020001@p2-29.van.tvs.net> Delivery-Date: Sat, 06 Sep 1997 01:31:33 -0700 Return-Path: news@crcnis3.unl.edu for <gherbert@crl.com>; Sat, 6 Sep 1997 01:01:06 -0700 (PDT) env-from (news@crcnis3.unl.edu) id QQdfum21077; Sat, 6 Sep 1997 04:05:47 -0400 (EDT) To: sci-space-tech@uunet.uu.net Alex Curylo (acurylo@inmediapresents.nospam.com) wrote: : : The PowerPC family dominates a large number of embedded systems markets. : Besides cars, there's a PPC in pretty much every piece of medical imaging : hardware, for instance. Not to mention the Mars Rover. (PPC -- 100% of the : Martian computing installed base! There's an ad for you huh?) : Not to rain on your extraterrestial parade or anything, but doesn't the Mars Rover use a version of the Intel 8085? I could almost swear I saw it in _Aviation Week_ or someplace. -Josh
From: Mirko Viviani <mirko@procom.it> Newsgroups: comp.sys.next.programmer Subject: Library problems... Date: 6 Sep 1997 08:49:14 GMT Organization: GMS & EMS development. Message-ID: <5ur5ea$apa@next.procom.it> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Ciao... I have two problems with a linker library and a rld lib under NS 3.3... 1. I need to create a linker library (libXX.a) for my own... well I put all my object files to this library with libtool but due to long file names >16 chars it truncates them ! Is there a way to put files >16 chr to a linker library ? 2. I want to use a dll for a my prg using rld functions but I have a problem... This rld need functions from NeXT system library like libNeXT_s.a and the only method I know to use them from my rld is to link them to the exe with -all_load but this causes an huge exe. Is there a way to include these functions to my rld lib w/o linking them to the exe ? Thank you... -- Bye, Mirko <mirko@procom.it> (NeXTmail, MIME)
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Objective-C Fragile Base Class (long) Date: 07 Sep 1997 13:31:23 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7bu257438.fsf@tom.ics.ele.tue.nl> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> <5umnvq$ju9$1@gte2.gte.net> <34100283.0@epflnews.epfl.ch> <5uph79$pvu$1@gte1.gte.net> <x7u3fy2ul2.fsf@tom.ics.ele.tue.nl> <5usc87$2fq$1@gte2.gte.net> In-reply-to: ga@ed4u.com's message of 6 Sep 1997 19:51:35 GMT In article <5usc87$2fq$1@gte2.gte.net> ga@ed4u.com (G. Apple) writes: Often, MI does provide a way to extend someone else's library through overrides and MI extension -- but yes, it does often require that they cooperate by making the necessary methods virtual. It also requires that the library does not allocate objects: If it provides a class A that does not really suit your application, and you create a subclass, that subclass is of no use if the library allocates instances of A itself. --Tiggr
From: d89cb@efd.lth.se (Christian Brunschen) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 7 Sep 1997 11:48:34 GMT Organization: Lund Institute of Technology, Sweden Message-ID: <5uu4ai$47v$1@news.lth.se> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> NNTP-Posting-User: d89cb In article <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com>, Daniel L. Taylor <LTaylor7@ix.netcom.com> wrote: >In article <5upjgt$sme$1@news.lth.se>, d89cb@efd.lth.se (Christian >Brunschen) wrote: > >> >> Ah, I get it -- the letters 'NC' scare you. Well, a 'NC' need not be >> slow, RAMless, or diskless. > >I have yet to see a NC spec which called for much more. You tried to >refute me in your post by straddling the line between NC and PC, missing >the point that PC will win out easily no matter how closely you straddle >that line. You're the one who doesn't get it: There is no such 'line'. An 'NC' is, basically, a PC that is administered centrally, > >> >Imagine! Everything that's yours stored somewhere else! No privacy! Web >> >like wait times for your applications and files! A dead computer when the >> >network goes down! >> >> Oh Please. Lots of people live quite comfortably with having their home >> directories mounted over a LAN, using, say, NFS (under Unix). If you start >> putting your servers and clients far apart, connected through 28k8 modems, >> of course performancs will suffer. But put them on a 100Mb/s Ethernet, and >> I doubt anyone will have any complaints -- especially since people will be >> able to choose any one of a bunch of identical computers and get their own >> personal computing environment. > >Depends on who you are and what you're doing. Believe me, plenty of people >would complain and loudly. Yes. And if you're doing something that requires lots of disk accesses etc, fetch the file to the local cesondary storage, run your program there, and then put the result back. You know, the right tool for the right job and all that. There will always be instances of problems where a certain solution doesn't apply. That does not, however, invalidate that solution for _other_ problems. > >And what advantage will any NC offer here over a PC? None. Ease of administration. A 'PC', by definition, is a _Personal Computer_ -- ie, one which is intended for one person's use. An NC differs from a PC not so much in hardware, as in how it is administrated. You seem to fail to graps that. > >> >It's the future people. A graphic designer won't mind waiting while >> >his/her NC swaps MB after MB of program and file data back and forth over >> >an Ethernet line to apply a single filter because MIS will save so much >> >money on support costs. They will love NC's and jump right on them to save >> >in support costs, just like they jumped right on Macintosh to save in >> >support costs, right??? >> >> It's about adminitering all machines centrally. If all machines are NC:s - >> in the sense that they aren't 'individual' machines, but rather 'clones' >> regarding OS, software, and even most hardware - then you can shave off all >> support costs regarding support of the 'individual' machines, since all >> machines will be interchangable. Plus, you will be able to have a few 'fresh' >> machines in 'reserve', so if any machine in the 'field' breaks down, you can >> swap it for one that works, without further interruption to the work of >> whoever is using the machine in question. > >And what advantage will NC's offer here over PC's? Are you seriously >suggesting that PC's cannot be configured in the same way, especially as >networking tools become more advanced? But if you configure each coputer in the network identically, and administer them centrally -- then you no longer have 'PC's (_Personal_ computers), but 'NC's (_Network_ computers). You migh have hardware that originally stems from a 'PC, but the computer will be a _Network_ Computer. Just like, if you take a brand-X Wintel clone and run Solaris on it, or OPENSTEP, it will be -- not a 'Personal Computer', but a Workstation. The _hardware_ makes little difference. > >Again, how is Apple going to succeed here? By providing a system that can be used standalone (like current Personal Computers), or within a network (as an NC -- administered centrally), while having a state-of-the-art user interface, best-of-breed applications, and best-of-breed development tools. > >> >They will gladly dump all their hardware just to be able to run slow Java >> >apps on a slow NC...right? Or dump native apps in favor of Java >> >interpreted apps served over a bogged down Ethernet line, right? Whoops! >> >Some repair guy just smashed through a wall and the networking cable. >> >Everybody go home now.... >> >> Um, Rhapsody runs, first and foremost, compiled Objective-C applications. >> JIT-compiled Java is hardly 'interpreted' either, btw. > >It's also hardly as fast. JIT-compiled java is a lot faster than 'interpreted' thank you. I don't have the benchmarks handy, but I _beleive_ (may be wrong, though) that Just-in-time compiles Java is, in speed, within an order of magnitude of compiled C++-code. And for _many_ applications, this is quite sufficient, especially given the 'write once, run everywhere' advantages of Java, etc. Not for all, obviously -- but if you have special needs, noone is going to force you to run on an NC, are they ? > >> >And consumers will flock to this new technology. Imagine loading Word at >> >28.8K. Heck, why doesn't AOL jump on this? All they have to do is >> >integrate a word processor, spreadsheet, and DB into the AOL client >> >software. Don't all of you want to do your work via AOL's server and store >> >your data there? If that picture doesn't send chills up and down your >> >spine, send your resume to Apple. They need people like you with "vision". >> >> You have heard the expression 'FUD', perhaps? Stands for 'Fear, Uncertainty, >> and Doubt'. This is usually backed up with baseless assertions, sometimes >> even outright lies. >> >> Just like you have been writing. > >Show me a spec which defines a full powered NC that is simply administered >from a server...and I'll show you a Windows PC "administered" by Bill >Gates. Well, sorry to burst your bubble, but if you place a bunch of Windows machines in a network, make them configurable over the net, and run centrally provided applications etc ... you have NC:s. Just like if you place a bunch of Sun workstations in a net, with central servers for such things as accounts, home directories, network addresses ... But these NC:s can still be fast, have lots of RAM, have lots of secondary storage ... Also, where is the original spec that defines that a PC should have umpteen Megabytes of RAM, a foo-hundred MHz CPU, a N-gigabyte harddisk, millions of pixels with milions of colours per pixel graphics .... Just because some people think that one version of the 'NC' is going to be slow, have little RAM, and no disk, doesn't mean that everybody thinks so, or even that that applies to all 'NC's. For some types of work (like simple word processing perhaps?) a slow CPU, little RAM, and no HD may well be quite sufficient. Why should anyone wast mney on something that isn't needed ? > >Steve has seen his vision...and he's going to watch as Gates passes it >right over. I don't think you are privy to Steve Jobs' visions any more than anyone else on this newsgroup. > >-- >Daniel L. Taylor >LTaylor7@ix.netcom.com >================================================================= >The Heritage Foundation /// Town Hall >http://www.heritage.org/heritage/ /// http://www.townhall.com/ >================================================================= // Christian Brunschen
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sun, 07 Sep 1997 06:14:46 -0700 Organization: Netcom Message-ID: <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> In article <5uu4ai$47v$1@news.lth.se>, d89cb@efd.lth.se (Christian Brunschen) wrote: > >I have yet to see a NC spec which called for much more. You tried to > >refute me in your post by straddling the line between NC and PC, missing > >the point that PC will win out easily no matter how closely you straddle > >that line. > > You're the one who doesn't get it: There is no such 'line'. An 'NC' > is, basically, a PC that is administered centrally, Not according to the specs I've seen. But if you do build a NC in such a fashion, what advantage does it have over the PC? That's what you are *not* getting. In other words: Why on Earth would any MIS department go with your NC over a Windows PC that they know, trust, and that has access not only to centralized administration but also tens of thousands of applications and tools, some of which they already know and own? Straddle the line that close to a PC and you have the same costs, which means your NC has *no* benefit *what-so-ever* over a standard Wintel PC. Go the opposite direction, towards a "true" NC (i.e. diskless, small RAM, cheap Java station) and you will have a market due to the low price, but it will be a small market because the NC simply will not be able to perform the same tasks as the "client heavy" PC. I get *really* irritated with NC and Java fans because they always respond with the "You just don't get it" line. "You don't understand...if you did, you would see it's the future." If you don't "see the light", you're a child that needs to be taught. I *do* understand because I've spent a *great* deal of time thinking about Java and NC's. And they're not a paradigm shift or "...the future..." as they are made out to be. They have their place and they will turn a profit in their markets for those who venture there. But they *are not* going to replace the "client heavy" PC. The idea of centralized administration has already been done, and it will improve on the PC to incorporate NC ideas, blurring the line between the two, but ultimately leaving "client heavy" PC's well entrenched. NC's represent an evolution of current networking ideas, not a paradigm shift that will cast Apple up into the clouds and Microsoft into the pit of dispair and/or radically change how we compute. As to Apple: when MacRumors first reported that Apple was started development on a NC, I approved. *Adding* NC's to Amelio's original plan is a good idea. Offer one highly scalable, PnP architecture (Rhapsody on CHRP) to fill the needs of high end server, workstation, PC, and NC markets. One solution, one API, yet it covers each base better than the other mix-n-match stuff out there. But *replacing* Amelio's original plan with NC's and abandoning the desktop and Apple's user base in the process ("...milk the Macintosh for all it's worth and move to the next great thing...") is damned foolishness and will get Apple buried. > >> >Imagine! Everything that's yours stored somewhere else! No privacy! Web > >> >like wait times for your applications and files! A dead computer when the > >> >network goes down! > >> > >> Oh Please. Lots of people live quite comfortably with having their home > >> directories mounted over a LAN, using, say, NFS (under Unix). If you start > >> putting your servers and clients far apart, connected through 28k8 modems, > >> of course performancs will suffer. But put them on a 100Mb/s Ethernet, and > >> I doubt anyone will have any complaints -- especially since people will be > >> able to choose any one of a bunch of identical computers and get their own > >> personal computing environment. > > > >Depends on who you are and what you're doing. Believe me, plenty of people > >would complain and loudly. > > Yes. And if you're doing something that requires lots of disk accesses > etc, fetch the file to the local cesondary storage, run your program > there, and then put the result back. This isn't going to work for a number of applications. Even with 100 mega*bit* per second networks. Ethernet is just now achieving speeds of fast dedicated (i.e. 1 computer accessing the disk) SCSI buses, which are already considered too slow. > You know, the right tool for the right job and all that. > > There will always be instances of problems where a certain solution > doesn't apply. That does not, however, invalidate that solution for > _other_ problems. No one ever said it did. But the mocking picture I painted of NC's is exactly how they will look to the current Mac base (or most of it any way). Apple cannot abandon their current Mac base by screwing IBM and Motorola, screwing the cloners, and screwing CHRP and hope to even survive long enough to build any kind of market in NC's. > >And what advantage will any NC offer here over a PC? None. > > Ease of administration. > > A 'PC', by definition, is a _Personal Computer_ -- ie, one which is > intended for one person's use. An NC differs from a PC not so much in > hardware, as in how it is administrated. You seem to fail to graps that. No, I grasp it perfectly. What you fail to grasp is that there is nothing preventing MS and Intel, and 3rd party developers, from making Wintel PC's about as easy to administer. If that's even a large factor. Look at how well this argument worked for Mac OS in corporations. Which leaves you with two choices: load your NC so it has all the power of a PC and lose out because the PC is just as good plus it runs Wintel software. Or go the more traditional NC route and build a very low cost diskless station that is preferred in some markets. Or be one of the 3rd party developers who evolve PC networking. Not exactly a paradigm shift there.... At most we can hope the competition adds a few impressive features to WinNT/WinXX. But which will Apple choose? And how will it succeed? > >And what advantage will NC's offer here over PC's? Are you seriously > >suggesting that PC's cannot be configured in the same way, especially as > >networking tools become more advanced? > > But if you configure each coputer in the network identically, and > administer them centrally -- then you no longer have 'PC's (_Personal_ > computers), but 'NC's (_Network_ computers). Again: which are you going to buy, the Wintel PC turned NC by a wave of a semantic wand, or an Apple branded NC? Apple's dead if they pursue this route to the exclusion of all else. I love the identical configuration pipe-dream to. This may work in some spots in the real world, but not all. Different workers are going to have different needs. The configurations are not going to be so "identical" as many suspect and you will still be subject to individual configuration problems, whether the user's private directory is a local HD (as now) or a server directory (NC). > >Again, how is Apple going to succeed here? > > By providing a system that can be used standalone (like current > Personal Computers), or within a network (as an NC -- administered > centrally), while having a state-of-the-art user interface, > best-of-breed applications, and best-of-breed development tools. Apple appears to be abandoning the idea of anything that's "standalone". The desktop wars are over and cloning/CHRP is dead, remember? Best UI I can buy. But best of breed applications and development tools? I think they are migrating as we speak to Wintel.... > >> >They will gladly dump all their hardware just to be able to run slow Java > >> >apps on a slow NC...right? Or dump native apps in favor of Java > >> >interpreted apps served over a bogged down Ethernet line, right? Whoops! > >> >Some repair guy just smashed through a wall and the networking cable. > >> >Everybody go home now.... > >> > >> Um, Rhapsody runs, first and foremost, compiled Objective-C applications. > >> JIT-compiled Java is hardly 'interpreted' either, btw. > > > >It's also hardly as fast. > > JIT-compiled java is a lot faster than 'interpreted' thank you. I > don't have the benchmarks handy, but I _beleive_ (may be wrong, > though) that Just-in-time compiles Java is, in speed, within an order > of magnitude of compiled C++-code. Java fans *never* have the benchmarks handy. Last I checked JIT was quite a bit slower than compiled C, and it was believed that even *compiled* Java would always be slower. But you are right, for many applications JIT will be fast enough, and the speed issue is not a major one for said applications. So again, why are corporations going to choose Apple Rhapsody/Java NC's over Microsoft Windows/Java "NC's"? What corporation trusts Apple at this point? They have killed every alliance and project they have proposed as the future. Even *I* don't trust them any more. Who is to say Apple won't turn around next year and abandon Rhapsody NC's in favor of something else? After all, if they are willing to screw their own partners, who is to say they won't screw you? That's the big picture you are missing. Even if NC's are the next insanely great thing, they will not be bought by corporations from Apple Computer. Not now.... > >Show me a spec which defines a full powered NC that is simply administered > >from a server...and I'll show you a Windows PC "administered" by Bill > >Gates. > > Well, sorry to burst your bubble, but if you place a bunch of Windows > machines in a network, make them configurable over the net, and run > centrally provided applications etc ... you have NC:s. Just like if > you place a bunch of Sun workstations in a net, with central servers > for such things as accounts, home directories, network addresses > ... But these NC:s can still be fast, have lots of RAM, have lots of > secondary storage ... And sorry to burst your bubble, but who is going to buy Apple's version over Microsoft's? > >Steve has seen his vision...and he's going to watch as Gates passes it > >right over. > > I don't think you are privy to Steve Jobs' visions any more than > anyone else on this newsgroup. Sure I am. Just read the man's quotes about milking the Mac market and abandoning the desktop for the "next great thing".... -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: d89cb@efd.lth.se (Christian Brunschen) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 7 Sep 1997 18:19:30 GMT Organization: Lund Institute of Technology, Sweden Message-ID: <5uur7i$9po$1@news.lth.se> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> NNTP-Posting-User: d89cb In article <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com>, Daniel L. Taylor <LTaylor7@ix.netcom.com> wrote: > >In article <5uu4ai$47v$1@news.lth.se>, d89cb@efd.lth.se (Christian >Brunschen) wrote: > >> >I have yet to see a NC spec which called for much more. You tried to >> >refute me in your post by straddling the line between NC and PC, missing >> >the point that PC will win out easily no matter how closely you straddle >> >that line. >> >> You're the one who doesn't get it: There is no such 'line'. An 'NC' >> is, basically, a PC that is administered centrally, > >Not according to the specs I've seen. Well, whatever specs you have seen, probably define what _ine_ person, or company, saw as 'the NC' -- ie, a product they want to build and want to create some sort of hype for; but I can't accept that that definition would define what NC:s are in general. > >But if you do build a NC in such a fashion, what advantage does it have >over the PC? That's what you are *not* getting. In other words: Why on >Earth would any MIS department go with your NC over a Windows PC that they >know, trust, and that has access not only to centralized administration >but also tens of thousands of applications and tools, some of which they >already know and own? Perhaps because the applications they want to run exist on the other available platform(s), too, and those are easier to a) use, b) administrate ? > >Straddle the line that close to a PC and you have the same costs, which >means your NC has *no* benefit *what-so-ever* over a standard Wintel PC. Oh yes -- the software: the User interface, the Administration software, the underlying OS .... > >Go the opposite direction, towards a "true" NC (i.e. diskless, small RAM, >cheap Java station) and you will have a market due to the low price, but >it will be a small market because the NC simply will not be able to >perform the same tasks as the "client heavy" PC. Well, in some markets those tasks are not about to be performed. A 'terminal' in a book shop, connected to a database of books with descriptions, ISB-numbers (or ISBN:s -- but not 'ISBN numbers' please :), will never need the kind of crunching power that lets you run Quake. > >I get *really* irritated with NC and Java fans because they always respond >with the "You just don't get it" line. "You don't understand...if you did, >you would see it's the future." If you don't "see the light", you're a >child that needs to be taught. (Just to set one thing straight: I am not a 'NC and Java fan' in the sense you seem to paint them here -- I 'envision' an 'NC' in the sense of a rather heavily loaded, hardware-wise, Rhapsody machine (be it Intel-, PPC-, Alpha-, SPARC-, MIPS-, PA-RISC- or whatever-based), which is centrally administered, gets its software 'pushed' to ir from a central server, and which mounts users' home directories remote, but which has enough secondary storage that jobs that need to do lots of stuff with lots of data on dis, can do that locally. Nothing that OPENSTEP machines don't offer you today, actually -- but by using the 'NC' moniker, you get people to actually _do_ that, instead of setting each machine up as a 'Personal Workstation' and just connecting them loosely to a network.) I think that the Java-based NC concept is a fine _concept_, and that very solution will probably be used in some places. But, obviously, not everywhere -- nor should it. > >I *do* understand because I've spent a *great* deal of time thinking about >Java and NC's. And they're not a paradigm shift or "...the future..." as >they are made out to be. They have their place and they will turn a profit >in their markets for those who venture there. We are in complete agreement here. The 'paradigm shift' is basically in the minds of people who have _not_ been doing these things already (as many people have, though under other names than 'NC'). >But they *are not* going to >replace the "client heavy" PC. The idea of centralized administration has >already been done, and it will improve on the PC to incorporate NC ideas, >blurring the line between the two, but ultimately leaving "client heavy" >PC's well entrenched. NC's represent an evolution of current networking >ideas, not a paradigm shift that will cast Apple up into the clouds and >Microsoft into the pit of dispair and/or radically change how we compute. > >As to Apple: when MacRumors first reported that Apple was started >development on a NC, I approved. *Adding* NC's to Amelio's original plan >is a good idea. Offer one highly scalable, PnP architecture (Rhapsody on >CHRP) to fill the needs of high end server, workstation, PC, and NC >markets. One solution, one API, yet it covers each base better than the >other mix-n-match stuff out there. > >But *replacing* Amelio's original plan with NC's and abandoning the >desktop and Apple's user base in the process ("...milk the Macintosh for >all it's worth and move to the next great thing...") is damned foolishness >and will get Apple buried. Well, Apple's development of Rhapsody looks arather like they will be putting their money not with the 'Java-based, underpowered NC's you seem to think of, but rather in the area of 'centrally administered, powerfol workstation-class machines' that I envision :) OPENSTEP itself is somewhat resource-hungry, so by necessity a Rhapsody-based 'NC' would have to have sufficient resources to run both Rhapsody itself, and whatever applications the users want to run on top of this. >> >> Yes. And if you're doing something that requires lots of disk accesses >> etc, fetch the file to the local cesondary storage, run your program >> there, and then put the result back. > >This isn't going to work for a number of applications. Even with 100 >mega*bit* per second networks. Ethernet is just now achieving speeds of >fast dedicated (i.e. 1 computer accessing the disk) SCSI buses, which are >already considered too slow. Well, in those settings, use a tool better suited to the task, or arrange the task so that you can perform it efficiently with your tools! Nothing says that you can't have a huge computing&disk server in the middle of your NC:s onto which you can lay all your heavy jobs, while running the (also computing-intenseive, perhaps) GUI frontend on your 'NC'. > >> You know, the right tool for the right job and all that. >> >> There will always be instances of problems where a certain solution >> doesn't apply. That does not, however, invalidate that solution for >> _other_ problems. > >No one ever said it did. But the mocking picture I painted of NC's is >exactly how they will look to the current Mac base (or most of it any >way). Apple cannot abandon their current Mac base by screwing IBM and >Motorola, screwing the cloners, and screwing CHRP and hope to even survive >long enough to build any kind of market in NC's. In another post of mine I remarked that one should not read too harsh 'rejection' into the words 'We have no plans to', which were what CFO Anderson said about a few things -- such as, CHRP licensing, and licensing MacOS 8. Interestingly, no less than 3 cloners have announced shipping boxes with MacOS 8, or acquiring a license to do so -- UMAX, PowerTools and .. Power Computing. Sort of corroborates my guess that 'We have no plans to ...' meant, generally, 'no comment, but here's a phrase you can print and feel like you have an answer'. > >> >And what advantage will any NC offer here over a PC? None. >> >> Ease of administration. >> >> A 'PC', by definition, is a _Personal Computer_ -- ie, one which is >> intended for one person's use. An NC differs from a PC not so much in >> hardware, as in how it is administrated. You seem to fail to graps that. > >No, I grasp it perfectly. What you fail to grasp is that there is nothing >preventing MS and Intel, and 3rd party developers, from making Wintel PC's >about as easy to administer. If that's even a large factor. Look at how >well this argument worked for Mac OS in corporations. Oh, there is -- Microsoft Windows :) In _theory_, nothing prevents anyone from doing anything with any OS. Yet often users find that certain tasks are more easily accomplished in one environment, compared to others. You know, one of the reasons why people, and companies, choose Macintoshen instead of Wintel boxes in the first place. > >Which leaves you with two choices: load your NC so it has all the power of >a PC and lose out because the PC is just as good plus it runs Wintel >software. Or go the more traditional NC route and build a very low cost >diskless station that is preferred in some markets. Or be one of the 3rd >party developers who evolve PC networking. Or, of course, offer, simply, a solution that works better, and can thus save you money. Yes, there's lots of applications for Wintel that won't run natively on Mac. So ? Virtual PC. Put that in your Rhapsody-based NCs, and you can run your legacy Wintel Apps just fine. > >Not exactly a paradigm shift there.... At most we can hope the competition >adds a few impressive features to WinNT/WinXX. > >But which will Apple choose? And how will it succeed? Apple will produce a system that gives you a powerful, stable OS, with a rich yet easy-to-use and _consistent_ Graphical User Interface, powerful inter-application and inter-computer communication (Distributed Objects), standardized, powerful and easy-to-use Inter-Application and inter-computer scripting, and a Unix command line if you want it. Even if we compare today's Windows offerings (95 and NT) against today's 'Rhapsody-equivalent' (OPENSTEP 4.2), > >> >And what advantage will NC's offer here over PC's? Are you seriously >> >suggesting that PC's cannot be configured in the same way, especially as >> >networking tools become more advanced? >> >> But if you configure each coputer in the network identically, and >> administer them centrally -- then you no longer have 'PC's (_Personal_ >> computers), but 'NC's (_Network_ computers). > >Again: which are you going to buy, the Wintel PC turned NC by a wave of a >semantic wand, or an Apple branded NC? Since I firmly beleive that Apple's solution will be better than that offered by Wintel, I would purchase an 'Apple branded NC'. And I further beleive that enough people and companies besides me will come to similar conclusions. Sufficiently many, even, to keep Apple a viable company, and Rhapsody a viable platform. > >Apple's dead if they pursue this route to the exclusion of all else. Since Apple are not betting their money on the 'underpowered java-runtime in hardware' as far as I can see, I must disagree with this statement. > >I love the identical configuration pipe-dream to. This may work in some >spots in the real world, but not all. As with everything, there is no one solution that fits everyone. >Different workers are going to have >different needs. The configurations are not going to be so "identical" as >many suspect and you will still be subject to individual configuration >problems, whether the user's private directory is a local HD (as now) or a >server directory (NC). The problem isn't the user's (mis)configuration of his own working environment -- that cannot be 'gotten around' except by disallowing (re)configuration your own environment (this may of course be a viable solution in many circumstances). What you gain by putting the _machine's_ configuration in a central place is such things as, ease of network trouble shooting et al. > >> >Again, how is Apple going to succeed here? >> >> By providing a system that can be used standalone (like current >> Personal Computers), or within a network (as an NC -- administered >> centrally), while having a state-of-the-art user interface, >> best-of-breed applications, and best-of-breed development tools. > >Apple appears to be abandoning the idea of anything that's "standalone". >The desktop wars are over and cloning/CHRP is dead, remember? A network of one machine, which is its own client and server, can be viewed as a 'stand-alone' confugration :) > >Best UI I can buy. But best of breed applications and development tools? I >think they are migrating as we speak to Wintel.... The Rhapsody development tools are available on WinNT, yes. But they still integrate best, and leverage from the system, under OPENSTEP / Rhapsody. Holism works -- The whole is greater than the sum of its parts. >> >> JIT-compiled java is a lot faster than 'interpreted' thank you. I >> don't have the benchmarks handy, but I _beleive_ (may be wrong, >> though) that Just-in-time compiles Java is, in speed, within an order >> of magnitude of compiled C++-code. > >Java fans *never* have the benchmarks handy. Last I checked JIT was quite >a bit slower than compiled C, and it was believed that even *compiled* >Java would always be slower. Well, you should be comparing Java to C++ or Objective-C. And in those cases, Java's performance is not so catastrophic as you paint it to be. And again, I am not a great Java advocate -- I prefer Objective-C :) > >But you are right, for many applications JIT will be fast enough, and the >speed issue is not a major one for said applications. (Even 'interpreted' Java will be fast enough for some applications ;) > >So again, why are corporations going to choose Apple Rhapsody/Java NC's >over Microsoft Windows/Java "NC's"? What corporation trusts Apple at this >point? They have killed every alliance and project they have proposed as >the future. Even *I* don't trust them any more. 'Even' you ? Considering there are still lots of people who do trust Apple enough to want to buy their systems (hardware and software), your threshold may not have been as high as you thought. A few of the larger stock analysts seem to trust Apple well enough to advise people to buy, rather than sell, Apple stock. Clone makers like PowerTools and UMAX seem to trust them enough to ship Mac clones with MacOS 8. > >Who is to say Apple won't turn around next year and abandon Rhapsody NC's >in favor of something else? After all, if they are willing to screw their >own partners, who is to say they won't screw you? > >That's the big picture you are missing. Even if NC's are the next insanely >great thing, they will not be bought by corporations from Apple Computer. >Not now.... I disagree on several points. Not everyone distrusts Apple the way you claim they do -- opinions differ on the issues; some people, including well-renowned indostry analists, consider Apple's recent moves to be good moves -- for Apple. > >> >Show me a spec which defines a full powered NC that is simply administered >> >from a server...and I'll show you a Windows PC "administered" by Bill >> >Gates. >> >> Well, sorry to burst your bubble, but if you place a bunch of Windows >> machines in a network, make them configurable over the net, and run >> centrally provided applications etc ... you have NC:s. Just like if >> you place a bunch of Sun workstations in a net, with central servers >> for such things as accounts, home directories, network addresses >> ... But these NC:s can still be fast, have lots of RAM, have lots of >> secondary storage ... > >And sorry to burst your bubble, but who is going to buy Apple's version >over Microsoft's? The same people who currently buy Apple's products now over Microsofts -- those who buy what they need rather than what they are told they need ? :) > >> I don't think you are privy to Steve Jobs' visions any more than >> anyone else on this newsgroup. > >Sure I am. Just read the man's quotes about milking the Mac market and >abandoning the desktop for the "next great thing".... And what makes you think that Steve is necessarily telling all his visions ? > >-- >Daniel L. Taylor >LTaylor7@ix.netcom.com >================================================================= >The Heritage Foundation /// Town Hall >http://www.heritage.org/heritage/ /// http://www.townhall.com/ >================================================================= // Christian Brunschen
From: mmalcolm crawford <malcolm@plsys.co.uk> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 7 Sep 1997 18:06:33 GMT Organization: P&L Systems Message-ID: <5uuqf9$4n4$1@ironhorse.plsys.co.uk> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> <5utdrs$41l$1@attila.apana.org.au> In-Reply-To: <5utdrs$41l$1@attila.apana.org.au> On 09/07/97, Matt McLeod wrote: >Still, with Rhapsody being reasonably cross-platform, Apple could get away >with sourcing PPC parts from IBM only, as they'd have the option of >producing systems with other CPUs (so IBM PPC parts would be competing with >Intel, Cyrix, AMD, IBM x86, maybe even Sparc and Alpha - if the Linux >community can cover those platforms, then surely a multi-billion-dollar >corporation could manage it?). > >Maybe Apple just need to hire Linus Torvalds to run their OS-development. > Why should they, when they have Avie Tevanian, who had NEXTSTEP running on m68k, x86, HP/PA-RISC, SPARC, PPC and probably a few others. Best wishes, mmalc.
From: mtrombin@ix.netcom.com (Mark Trombino) Newsgroups: comp.sys.next.programmer Subject: NSWindow hangs in new thread Date: 7 Sep 1997 18:32:18 GMT Organization: Egghead Billy, Inc. Message-ID: <5uurvi$on5@dfw-ixnews12.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Howdy! Here's todays brain tickler.... :) Is there anything that would cause a window that was order to the front of the screen in one thread and then ordered off screen in a separate thread to hang in it's orderOut: method? Here's the problem... The user pushes a button. That button calls a method that orders a regular NSWindow to the front (non-modal) and then spawns a new thread that calls a method to initiate pppd: - (void)startPPPWithScript:(id)scriptString { int result; NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init]; result = system([scriptString cString]); if(result == 256) // Didn't establish pppd [self cleanUp]; [pool release]; [NSThread exit]; } So far so good. The problem occurs if pppd fails and the method cleanUp gets called. That method just resets everything back to a normal state. Part of that method ordersOut the NSWindow that was previously ordered to the front. - (void) cleanUp { [connectionTimer invalidate]; connectionTimer = nil; [pppCheckTimer invalidate]; pppCheckTimer = nil; [[self currentFromLocation] setPPPActive:NO]; [[self currentToLocation] setPPPActive:NO]; [[self currentOptionLocation] setPPPActive:NO]; [imageViewer setImage:offImageBack]; // Reset window to reflect unconnected state. [connectButton setTitle:LOGON_OFFLINEBUTTONTEXT]; [connectMenuButton setTitle:@"Connect NOW"]; [connectButton setTag:LOGON_STATEOFF]; [fromPopup setEnabled:YES]; [toPopup setEnabled:YES]; [fromOptionMenu setEnabled:YES]; [toOptionMenu setEnabled:YES]; [optionPopup setEnabled:YES]; [optionOptionMenu setEnabled:YES]; [connectButton setEnabled:YES]; [connectMenuButton setEnabled:YES]; // Restore /etc/ppp/pap-secrets if(restorePAPSecrets) { [[NSFileManager defaultManager] removeFileAtPath:LOGON_PAPSECRETSPATH handler:nil]; if([[NSFileManager defaultManager] fileExistsAtPath:LOGON_PAPSECRETSSAVEDPATH]) { [[NSFileManager defaultManager] movePath:LOGON_PAPSECRETSSAVEDPATH toPath:LOGON_PAPSECRETSPATH handler:nil]; } restorePAPSecrets = NO; } [messageWindow orderOut:self]; // <<<<<< PROBLEM METHOD!!! } The problem is that the thread hangs on the method orderOut. The window goes away but the method never returns! The original thread isn't affected at all, but the new thread is stuck, so I get a wait cursor, but the app continues to run. Before I made the app spawn a new thread to initiate pppd, I never had a problem with the cleanUp method. Can anyone see what I'm doing wrong that would cause this to happen? If I remove that problem line everything is fine (except the window is left on screen). What's going on here? Thanks! -- |-------------------------------------------------------------------------| | Mark Trombino | J A M S o f t | | mtrombin@ix.netcom.com | Audio DSP Tools for Openstep & Rhapsody | |-------------------------------------------------------------------------|
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sun, 07 Sep 1997 12:38:09 -0700 Organization: Netcom Message-ID: <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> In article <5uur7i$9po$1@news.lth.se>, d89cb@efd.lth.se (Christian Brunschen) wrote: > >> You're the one who doesn't get it: There is no such 'line'. An 'NC' > >> is, basically, a PC that is administered centrally, > > > >Not according to the specs I've seen. > > Well, whatever specs you have seen, probably define what _ine_ person, > or company, saw as 'the NC' -- ie, a product they want to build and > want to create some sort of hype for; but I can't accept that that > definition would define what NC:s are in general. NC's have been generally proposed as low-cost diskless stations which survive off of a server and lower the per-seat costs for corporate computing by lowering the initial cost in addition to administration costs (centralized setup, simplified Java-OS). I'm not denying that you can build a NC that has more power at the seat, but the idea of NC's is to move away from client-heavy computing. > >But if you do build a NC in such a fashion, what advantage does it have > >over the PC? That's what you are *not* getting. In other words: Why on > >Earth would any MIS department go with your NC over a Windows PC that they > >know, trust, and that has access not only to centralized administration > >but also tens of thousands of applications and tools, some of which they > >already know and own? > > Perhaps because the applications they want to run exist on the other > available platform(s), too, and those are easier to a) use, b) > administrate ? The applications don't run or exist on the other platform (hell, the other platform doesn't even yet exist) and you're assuming a) that they will be easier to use and administer and b) that it would matter if they really were. Mac OS is an order of magnitude easier to use and administer than Wintel and already has a jumpstart on applications, yet which dominates corporations? Know why? Basically for fears that Apple just managed to reinforce. > >Straddle the line that close to a PC and you have the same costs, which > >means your NC has *no* benefit *what-so-ever* over a standard Wintel PC. > > Oh yes -- the software: the User interface, the Administration > software, the underlying OS .... *What* software??? Mac OS has 12,000 applications, but how many are available for NC's? How many Java apps currently exist? OpenStep apps? What percentage of the Wintel library is mirrored by the Mac OS library, and what percentage by the Java library? The OpenStep library? Jobs is most likely going with an x86 NC. What apps will it have? Will it dual boot Windows? That's interesting...now it needs a local OS. If so, why not just use Wintel PCs? Ease of use and costs alone, eh? If ease of use and lower costs aren't enough to put Mac OS, with 12,000 applications, into corporations, what makes you think it's enough to put NC's after Apple has dropped the desktop, PowerPC, and screwed more partners? Under different conditions I would agree with you. But Apple has blown it's feet and lower legs out from under it with the recent move to close and abandon the Mac market. That puts any Apple NC at a major disadvantage. > (Just to set one thing straight: I am not a 'NC and Java fan' in the > sense you seem to paint them here -- I 'envision' an 'NC' in the sense > of a rather heavily loaded, hardware-wise, Rhapsody machine (be it > Intel-, PPC-, Alpha-, SPARC-, MIPS-, PA-RISC- or whatever-based), > which is centrally administered, gets its software 'pushed' to ir from > a central server, and which mounts users' home directories remote, but > which has enough secondary storage that jobs that need to do lots of > stuff with lots of data on dis, can do that locally. Nothing that > OPENSTEP machines don't offer you today, actually -- but by using the > 'NC' moniker, you get people to actually _do_ that, instead of setting > each machine up as a 'Personal Workstation' and just connecting them > loosely to a network.) This is *not* the traditionally defined NC. You are describing a "client heavy" PC, basically. Which might sell well except you're missing the point's that: * Apple just screwed more people in another round of "we promised this but will not deliver", shaking any faith anybody has in them even more. * We now do not know what Apple's NC will look like. Client heavy super machine or diskless station running what OS and capable of what? * Apple appears to be ditching both PowerPC and the desktop, losing four major advantages (Mac OS customer base, Mac OS compatibility - which won't mean anything without a customer base and growing software library, superior price/performance, and a scalable strategy that lets them service a wide range of needs). > >As to Apple: when MacRumors first reported that Apple was started > >development on a NC, I approved. *Adding* NC's to Amelio's original plan > >is a good idea. Offer one highly scalable, PnP architecture (Rhapsody on > >CHRP) to fill the needs of high end server, workstation, PC, and NC > >markets. One solution, one API, yet it covers each base better than the > >other mix-n-match stuff out there. > > > >But *replacing* Amelio's original plan with NC's and abandoning the > >desktop and Apple's user base in the process ("...milk the Macintosh for > >all it's worth and move to the next great thing...") is damned foolishness > >and will get Apple buried. > > Well, Apple's development of Rhapsody looks arather like they will be > putting their money not with the 'Java-based, underpowered NC's you > seem to think of, but rather in the area of 'centrally administered, > powerfol workstation-class machines' that I envision :) OPENSTEP > itself is somewhat resource-hungry, so by necessity a Rhapsody-based > 'NC' would have to have sufficient resources to run both Rhapsody > itself, and whatever applications the users want to run on top of > this. Amelio's plan has been scratched in favor of pure NC's. And we don't actually know that they will run Rhapsody. They may, rather, run Newton OS. My point isn't that NC's are bad, it's that ditching the Mac market and going for pure NC's will kill Apple. > >No one ever said it did. But the mocking picture I painted of NC's is > >exactly how they will look to the current Mac base (or most of it any > >way). Apple cannot abandon their current Mac base by screwing IBM and > >Motorola, screwing the cloners, and screwing CHRP and hope to even survive > >long enough to build any kind of market in NC's. > > In another post of mine I remarked that one should not read too harsh > 'rejection' into the words 'We have no plans to', which were what CFO > Anderson said about a few things -- such as, CHRP licensing, and > licensing MacOS 8. > > Interestingly, no less than 3 cloners have announced shipping boxes > with MacOS 8, or acquiring a license to do so -- UMAX, PowerTools and > .. Power Computing. Sort of corroborates my guess that 'We have no > plans to ...' meant, generally, 'no comment, but here's a phrase you > can print and feel like you have an answer'. Power Computing is being bought by Apple and there is no point in denying them OS 8. Power Tools may just be shipping the box...we don't know yet. UMAX just has OS 8, no CHRP. Motorola is screwed. There are no other cloners. This is *not* an open market and I think the Mac base has every right to read it harshly. > >No, I grasp it perfectly. What you fail to grasp is that there is nothing > >preventing MS and Intel, and 3rd party developers, from making Wintel PC's > >about as easy to administer. If that's even a large factor. Look at how > >well this argument worked for Mac OS in corporations. > > Oh, there is -- Microsoft Windows :) LOL! OK, so you made me eat that last statement. > In _theory_, nothing prevents anyone from doing anything with any > OS. Yet often users find that certain tasks are more easily > accomplished in one environment, compared to others. You know, one of > the reasons why people, and companies, choose Macintoshen instead of > Wintel boxes in the first place. But these 60 million users of 28 million machines may not stick around long enough to even see if Apple's new machines can fit their needs. > Even if we compare today's Windows offerings (95 and NT) against > today's 'Rhapsody-equivalent' (OPENSTEP 4.2), Yet OpenStep was all but dead when Apple bought it and Windows domiantes. Don't you get why I'm worried yet? Apple is making the same lame-brained business decisions all over again. Those decisions would kill even the computer off the Enterprise-E. Rhapsody and NC's don't stand a chance with business moves like these.... > >Again: which are you going to buy, the Wintel PC turned NC by a wave of a > >semantic wand, or an Apple branded NC? > > Since I firmly beleive that Apple's solution will be better than that > offered by Wintel, I would purchase an 'Apple branded NC'. And I > further beleive that enough people and companies besides me will come > to similar conclusions. Sufficiently many, even, to keep Apple a > viable company, and Rhapsody a viable platform. Will they? You're speaking in a newsgroup full of Mac faithful, who know how much better Apple's technology is, yet they are ready to leave due to business screw-ups. > >So again, why are corporations going to choose Apple Rhapsody/Java NC's > >over Microsoft Windows/Java "NC's"? What corporation trusts Apple at this > >point? They have killed every alliance and project they have proposed as > >the future. Even *I* don't trust them any more. > > 'Even' you ? Considering there are still lots of people who do trust > Apple enough to want to buy their systems (hardware and software), > your threshold may not have been as high as you thought. We will not know how "many" until the complete fallout from killing clones is felt, during the fall sales season. My guess is that there will not be very "many" left. > >Who is to say Apple won't turn around next year and abandon Rhapsody NC's > >in favor of something else? After all, if they are willing to screw their > >own partners, who is to say they won't screw you? > > > >That's the big picture you are missing. Even if NC's are the next insanely > >great thing, they will not be bought by corporations from Apple Computer. > >Not now.... > > I disagree on several points. Not everyone distrusts Apple the way you > claim they do -- opinions differ on the issues; some people, including > well-renowned indostry analists, consider Apple's recent moves to be > good moves -- for Apple. I see a lot of customers, Mac faithful, walking away now. Analysts can say what they will...they also said Apple should have died 10 years ago...if they are now positive about Apple, it really could be the end! > >And sorry to burst your bubble, but who is going to buy Apple's version > >over Microsoft's? > > The same people who currently buy Apple's products now over Microsofts > -- those who buy what they need rather than what they are told they > need ? :) The same ones voting against Apple in online polls and posting messages like "F*** you Jobs! I'm buying Wintel!".... -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
Date: Sun, 07 Sep 1997 14:24:38 -0700 From: gesNOSPAM@sisna.com (ges) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Message-ID: <gesNOSPAM-0709971424390001@dialup45.sisna.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <5up1dc$ndg$1@news.lth.se> <341074F1.8F26E657@spamtoNull.com> <5urae7$f7c$1@attila.apana.org.au> In article <5urae7$f7c$1@attila.apana.org.au>, mjm@attila.apana.org.au (Matt McLeod) wrote: > > This is what Rhapsody buys Apple - independence from Motorola. The Wintel > clone makers have had this for years now (able to buy from Intel, Cyrix, or > AMD). There's no real reason why they have to stop with x86, either - > Rhapsody could potentially support quite a lot of different CPUs, giving > Apple (and cloners, if they're able to continue) lots of options. > > Sounds pretty good to me. > > All this rationalizing doesn't reassure me. I think Apple is committing suicide, or rather Jobs is murdering it. gs
From: tbrown@netset.com (Ted Brown) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: Sun, 07 Sep 1997 11:21:14 -0500 Organization: All USENET -- http://www.Supernews.com Message-ID: <tbrown-0709971121140001@user038.axom.com> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> <mschemen-0409972046570001@ts005d05.ftw-tx.concentric.net> In article <mschemen-0409972046570001@ts005d05.ftw-tx.concentric.net>, mschemen@concentric.net (Mathew Schemenaur) wrote: >According to the expert Harry Norr, Apple was already getting 100 to 280 >per box when you concider software and hardware fees. So a jump to 100 to >500 does not seem like a jump at all. > >I think Gil got fired because he agreed to terms that were not much better >then the original terms. At least not enougth better to satisfy the board. That's what this CHRP mess is all about. Cloners don't have to pay Apple the MB fee ($180) for CHRP, that's a fairly big dent in the price of a $1000-$2000 machine. Apple is now facing the loss of that money, and is trying to stop it. It'd be nice to hear what Amelio actually agreed to, and what the pricing structure was for CHRP. I could see Apple charging more the CHRP lisence, as a CHRP machine need not include the Mac ROMs. The ROM contains code that is not in the MacOS. Thus a CHRP lisence is for the Mac OS and the Mac ROMs. It's a net win for Apple and the cloners as they don't have to expend the resources to make the ROM and put it on the mb. Long term, it's a better deal for Apple to allow CHRP to come thru (or something open like it), as it gets multiple vendors making chips, instead of Apple doing all the R+D. This brings Apple's costs down, and Apple can still compete with the cloners (or slowly bow out of the hw market). -- Ted Brown tbrown@netset.com
From: mpaque.spa-am@nospam.wco.com (Mike Paquette) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sun, 07 Sep 1997 22:54:30 GMT Organization: Electronics Service Unit No. 16 Message-ID: <34132450.2186844@news.wco.com> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> <5utdrs$41l$1@attila.apana.org.au> <5uuqf9$4n4$1@ironhorse.plsys.co.uk> On 7 Sep 1997 18:06:33 GMT, mmalcolm crawford <malcolm@plsys.co.uk> wrote: >Why should they, when they have Avie Tevanian, who had NEXTSTEP running on >m68k, x86, HP/PA-RISC, SPARC, PPC and probably a few others. I suspect that Avie had some help on that. Some of those people are now Apple employees, though. If an appreciable commercial market appears on other platforms, I suspect that Apple would target those platforms.
From: mpaque.spa-am@nospam.wco.com (Mike Paquette) Newsgroups: comp.sys.next.programmer Subject: Re: NSWindow hangs in new thread Date: Sun, 07 Sep 1997 22:54:42 GMT Organization: Electronics Service Unit No. 16 Message-ID: <34132b5f.3993642@news.wco.com> References: <5uurvi$on5@dfw-ixnews12.ix.netcom.com> On 7 Sep 1997 18:32:18 GMT, mtrombin@ix.netcom.com (Mark Trombino) wrote: >Howdy! > >Here's todays brain tickler.... :) > >Is there anything that would cause a window that was order to the front of >the screen in one thread and then ordered off screen in a separate thread to >hang in it's orderOut: method? Here's the problem... Urk! Well, I can try to point you in the right direction. Windows are owned by graphics contexts over in the WindowServer. Each context is associated with a NSDPSContext in your app, and there's a default DPS context per thread. It could be that the DPS context creating the window isn't the same as the one that's trying to terminate the window. I suggest that the easiest way to handle this might be to have the worker thread message back to your main event handling thread to close the window. You can use Distributed Objects to do this almost transparently. There are other more complex and fragile solutions, but this would be the first thing to try. Keeping graphics state on a single object synced between multiple threads and contexts can be tricky.
From: adt@netcom.com (Tony Tribelli) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sun, 07 Sep 1997 17:49:12 -0700 Organization: Delta Internet Services, Inc. Message-ID: <adt-ya023080000709971749120001@news.deltanet.com> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> <5utdrs$41l$1@attila.apana.org.au> <5uuqf9$4n4$1@ironhorse.plsys.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Matt McLeod wrote: > Maybe Apple just need to hire Linus Torvalds to run their OS-development. Why, his OS sales figures are far lower? :-) Tony ------------------ Tony Tribelli adtribelli@acm.org
From: ckane@next.com (Christopher Kane) Newsgroups: comp.sys.next.programmer Subject: Re: NSColor encoding for user defaults really missing ? Date: 8 Sep 1997 00:36:50 GMT Organization: Apple Computer, Inc. Distribution: world Message-ID: <5uvhb2$t72$1@news2.apple.com> References: <5um2in$s8p$1@concorde.ctp.com> In article <5um2in$s8p$1@concorde.ctp.com> tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) writes: > Thanx for all the replies which confirmed my suspicion that this > is really missing (and propably will not be in DR1 too) > > Don already has my NSColor (NSUserDefaults) category. What be > basically do is to use a string encoding that is pretty much > identical to: > [myColor description]; > > All you do is > > [myDefaults setColor:... forKey:...]; > [myDefaults colorForKey:...]; > > and we have the main NSColor additions: > > + (NSColor *)colorWithStringRepresentation:(NSString *)aString; > - (NSString *)stringRepresentation; The generic way of storing objects in NSUserDefaults that aren't property lists is to use an NSArchiver to archive the object into a data then an NSUnarchiver on read to reconstitute. This works for objects which conform to NSCoding. The methods: NSArchiver: + (NSData *)archivedDataWithRootObject:(id)rootObject; NSUnarchiver: + (id)unarchiveObjectWithData:(NSData *)data; make this fairly painless, and you don't have to define some string format (which could be slow to parse). Presumably archiving the objects provides the the greatest fidelity as well. Then tack a category onto NSUserDefaults, as Thomas did in his solution. Christopher Kane Application Frameworks Apple Computer, Inc.
From: ckane@next.com (Christopher Kane) Newsgroups: comp.sys.next.programmer Subject: Re: Stopping NSRunLoops Date: 8 Sep 1997 01:05:17 GMT Organization: Apple Computer, Inc. Distribution: world Message-ID: <5uvj0d$suq$1@news2.apple.com> References: <340C94E7.6BE2@ergotech.com> In article <340C94E7.6BE2@ergotech.com> Jim Redman <jim@ergotech.com> writes: > I feel that this should be a fairly dumb question, but just how do you > stop an NSRunLoop that has been started with a -run? Is it just in the > documentation and I missed it, or is there something clever about this? -run will keep iterating on an NSRunLoop as long as the run loop has sources of input, i.e. timers and ports. "ports" includes other things in Foundation too, which do background activities, such as reading from a file handle in the background or getting the exit notification for a spawned NSTask, which register ports with the appropriate run loop to trigger their notification in the invoking thread. Note that D.O. also works via ports. If you have control of all input sources (i.e. know all NSPorts that you've explicitly registered with the run loop, and know all D.O. connections, and know all timers, NSFileHandles, NSTasks, ...), you can invalidate or release these objects, which will remove their registered ports from the run loop as a side effect. Send -invalidate to all explicit NSPorts and NSConnections, remove timers explicitly from the run loop or cause them to stop otherwise, release retained NSFileHandles and NSTasks, etc. It might be simpler to run the run loop yourself and not use -run. Use -runUntilDate: or -runMode:beforeDate: in a loop which checks some external condition for when the thread should terminate, like: while (condition) { if (![[NSRunLoop currentRunLoop] runMode:NSDefaultRunLoopMode beforeDate:[NSDate dateWithTimeIntervalSinceNow:1.0]]) break; } This is essentially what -run is doing, except it uses [NSDate distantFuture]. Here, if -runMode:beforeDate: either return NO immediately, if the run loop has no input sources, or will block for up to 1 second waiting for something to happen (or less if something does happen). The condition could be some boolean flag with the thread knows about, for example. You don't want to make the condition too expensive to evaluate. Christopher Kane Application Frameworks Apple Computer, Inc.
Date: Sun, 07 Sep 1997 17:47:01 -0400 From: joe.ragosta@dol.net (Joe Ragosta) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Message-ID: <joe.ragosta-ya02408000R0709971747010001@news.dol.net> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> <5utdrs$41l$1@attila.apana.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit In article <5utdrs$41l$1@attila.apana.org.au>, mjm@attila.apana.org.au (Matt McLeod) wrote: > In article <B0372D47-F4AD@206.165.43.161>, > Lawson English <english@primenet.com> wrote: > : > :Nope. Ford Motor, and various other embedded market companies provide far > :greater markets for PPC's than APple's measely little 4 million/year. > > Thanks. I wasn't aware of this. That may be true on a volume basis. I doubt very much that it's true on a dollar basis. The average PPC that Apple buys is probably $300. That's $1 billion per year. The average PPC for Ford is probably $10-50. At 2-3 million per year, that's a tiny fraction of the revenues that Motorola gets from Apple. -- Regards, Joe Ragosta joe.ragosta@dol.net Visit the Complete Macintosh Web Site http://www.dol.net/~Ragosta/complmac.htm
From: mtrombin@ix.netcom.com (Mark Trombino) Newsgroups: comp.sys.next.programmer Subject: Re: NSWindow hangs in new thread Date: 8 Sep 1997 04:09:12 GMT Organization: Egghead Billy, Inc. Message-ID: <5uvtp8$qbh@dfw-ixnews11.ix.netcom.com> References: <5uurvi$on5@dfw-ixnews12.ix.netcom.com> <34132b5f.3993642@news.wco.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: mpaque@wco.com In <34132b5f.3993642@news.wco.com> Mike Paquette wrote: > On 7 Sep 1997 18:32:18 GMT, mtrombin@ix.netcom.com (Mark Trombino) > wrote: > > >Howdy! > > > >Here's todays brain tickler.... :) > > > >Is there anything that would cause a window that was order to the front of > >the screen in one thread and then ordered off screen in a separate thread to > >hang in it's orderOut: method? Here's the problem... > > Urk! > > Well, I can try to point you in the right direction. Windows are > owned by graphics contexts over in the WindowServer. Each context is > associated with a NSDPSContext in your app, and there's a default DPS > context per thread. > > It could be that the DPS context creating the window isn't the same as > the one that's trying to terminate the window. I suggest that the > easiest way to handle this might be to have the worker thread message > back to your main event handling thread to close the window. You can > use Distributed Objects to do this almost transparently. > > There are other more complex and fragile solutions, but this would be > the first thing to try. Keeping graphics state on a single object > synced between multiple threads and contexts can be tricky. > > Thanks Mike! I guess this is as good as time as any to start learning Distributed Objects! I had suspected something along those lines you mentioned, but hoped that there was a simple work around... Guess not :) Thanks for the pointer! -- |-------------------------------------------------------------------------| | Mark Trombino | J A M S o f t | | mtrombin@ix.netcom.com | Audio DSP Tools for Openstep & Rhapsody | |-------------------------------------------------------------------------|
From: jmeacham@meacham.charm.net (The Rev. James D. Meacham) Newsgroups: comp.lang.perl.misc,comp.sys.next.programmer,comp.sys.next.software Subject: mSQL mods for Perl compiled on NeXT? Date: 8 Sep 1997 04:24:50 GMT Organization: Charm.Net Baltimore Internet Access, Hon (410) 558-3900 Message-ID: <5uvumi$rp1$1@news2.charm.net> Hi all, I've been having a hell of a time trying to get the mSQL modules (version 1.1807) to compile on my NeXT 3.3 workstation. Anyone got these working. I've got all the requisite components, but it quits compiling at the MySQL section with complaints about undeclared variables. Any suggestions? Peace, James -- ------------------------------------------------------------------- The Rev. James David Meacham Minister, Webmaster, Philosopher "If people did not sometimes do silly things, nothing intelligent would ever get done." -- Ludwig Wittgenstein
From: tph@rmi.net (Tom "Tom" Harrington) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sun, 7 Sep 1997 22:44:59 -0600 Organization: Mechanist Industries Message-ID: <19970907224459112636@slip22126.rmii.com> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> <5utdrs$41l$1@attila.apana.org.au> Matt McLeod <mjm@attila.apana.org.au> wrote: > if the Linux > community can cover those platforms, then surely a multi-billion-dollar > corporation could manage it?). Sure, if they could get hundreds of skilled programmers to donate their time and energy the way Linus can. If the source code was free like it is with Linux, it could happen. And if I had wings, I'd be a bird. -- Tom Harrington ------- tph@rmii.com ------- http://rainbow.rmii.com/~tph "Rabbit's clever," said Pooh thoughtfully. "And he has Brain." "I suppose that's why he never understands anything." -A. A. Milne ----> CONSPIRACY ONLINE: http://rainbow.rmii.com/~tph/bunker.html <----
From: pohl@xxx.screaming.org (Pohl Longsine) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 05:36:01 GMT Organization: National Security Agency Message-ID: <slrn6174ok.cq1.pohl@xxx.screaming.org> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> <5utdrs$41l$1@attila.apana.org.au> <19970907224459112636@slip22126.rmii.com> Tom "Tom" Harrington <tph@rmi.net> wrote: >Matt McLeod <mjm@attila.apana.org.au> wrote: > >> if the Linux community can cover those platforms, then >> surely a multi-billion-dollar corporation could manage it?). > >Sure, if they could get hundreds of skilled programmers to >donate their time and energy the way Linus can. If the source >code was free like it is with Linux, it could happen. And if >I had wings, I'd be a bird. "...and if a frog had wings he wouldn't bump his ass a-hoppin'". There's a nice paper that formalizes this observation (about the linux development model -- not about the frog) by none other than the jargon-file's current maintainer Eric S. Raymond. http://www.ccil.org/~esr/writings/cathedral.html By the end of the article, the reader is almost ready to pity poor microsoft for having only a small fraction of the financial resources that would be required to scale-up debugging efforts to match those at work in some free-software projects like linux. There's some excellent material within.
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 15:54:59 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5v03vj$gkn$1@attila.apana.org.au> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> <5utdrs$41l$1@attila.apana.org.au> <5uuqf9$4n4$1@ironhorse.plsys.co.uk> In article <5uuqf9$4n4$1@ironhorse.plsys.co.uk>, mmalcolm crawford <malcolm@plsys.co.uk> wrote: : :On 09/07/97, Matt McLeod wrote: : :>Still, with Rhapsody being reasonably cross-platform, Apple could get away :>with sourcing PPC parts from IBM only, as they'd have the option of :>producing systems with other CPUs (so IBM PPC parts would be competing with :>Intel, Cyrix, AMD, IBM x86, maybe even Sparc and Alpha - if the Linux :>community can cover those platforms, then surely a multi-billion-dollar :>corporation could manage it?). :> :>Maybe Apple just need to hire Linus Torvalds to run their OS-development. :> :Why should they, when they have Avie Tevanian, who had NEXTSTEP running on :m68k, x86, HP/PA-RISC, SPARC, PPC and probably a few others. It was a "throw-away" line, not a serious suggestion. -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 15:58:29 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5v0465$guj$1@attila.apana.org.au> References: <5urae7$f7c$1@attila.apana.org.au> <5utdrs$41l$1@attila.apana.org.au> <5uuqf9$4n4$1@ironhorse.plsys.co.uk> <adt-ya023080000709971749120001@news.deltanet.com> In article <adt-ya023080000709971749120001@news.deltanet.com>, Tony Tribelli <adt@netcom.com> wrote: :Matt McLeod wrote: : :> Maybe Apple just need to hire Linus Torvalds to run their OS-development. : :Why, his OS sales figures are far lower? :-) Because the free Unix people (Linux, *BSD*, etc) have been able to support quite a few hardware platforms without the resources of a huge corporation behind them. If they can do that, imagine what they could do with a little money... -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 16:01:08 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5v04b4$h0a$1@attila.apana.org.au> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> <5utdrs$41l$1@attila.apana.org.au> <19970907224459112636@slip22126.rmii.com> In article <19970907224459112636@slip22126.rmii.com>, Tom "Tom" Harrington <tph@rmi.net> wrote: :Matt McLeod <mjm@attila.apana.org.au> wrote: : :> if the Linux :> community can cover those platforms, then surely a multi-billion-dollar :> corporation could manage it?). : :Sure, if they could get hundreds of skilled programmers to donate their :time and energy the way Linus can. If the source code was free like it :is with Linux, it could happen. And if I had wings, I'd be a bird. Then perhaps Apple (and other commercial software developers) should learn from the freeware experience. -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 16:04:42 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5v04hq$h1p$1@attila.apana.org.au> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <341074F1.8F26E657@spamtoNull.com> <5urae7$f7c$1@attila.apana.org.au> <gesNOSPAM-0709971424390001@dialup45.sisna.com> In article <gesNOSPAM-0709971424390001@dialup45.sisna.com>, ges <gesNOSPAM@sisna.com> wrote: :In article <5urae7$f7c$1@attila.apana.org.au>, mjm@attila.apana.org.au :(Matt McLeod) wrote: :> :> This is what Rhapsody buys Apple - independence from Motorola. The Wintel :> clone makers have had this for years now (able to buy from Intel, Cyrix, or :> AMD). There's no real reason why they have to stop with x86, either - :> Rhapsody could potentially support quite a lot of different CPUs, giving :> Apple (and cloners, if they're able to continue) lots of options. :> :> Sounds pretty good to me. : :All this rationalizing doesn't reassure me. I think Apple is committing :suicide, or rather Jobs is murdering it. Despite the popular view (which I share) that management is made up of idiots promoted beyond their competency, I don't really think that Apple is completely stupid. Killing themselves off is not conducive to long-term profitability. It looks to me like they've got some other "vision". Perhaps what they need to do now is articulate it. (And no, I don't beleive that it's this "NC" thing - at least, not the diskless, Java-based, thing). -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: pemmerik@solair1.inter.NL.net (P.J.L.van Emmerik) Newsgroups: comp.sys.next.programmer Subject: senderIsInvalid when working across networks Date: Mon, 08 Sep 1997 06:16:56 GMT Organization: NLnet Message-ID: <5v055s$ckn$1@news.NL.net> I have a client server situation. Clients connects to server using distributed objects. The problem is that the server receives a senderIsInvalid notification for clients on a different network, while those client apps are still very much alive, so a see no reason for the invalidation of the NXConnection. For clients on the local network (i.e. connected to router-B) all seems to work well. I am using NS 3.3 (intel) on all NEXT machines involved. Modem - high speed connection - modem +---------------------------------+ | | +--------+ +--------+ |Router-A| |Router-B| +--------+ +--------+ | | | | +------+ +------+ | |Client| |Server| | +------+ +------+ | Lots of other machines - can anyone give some pointers here? - how does sender invalidation work? Please Email to: emmereik@hpb.holec-projects.nl P.J.L. van Emmerik Holec Projects B.V. Email: emmerik@hpb.holec-projects.nl PO.BOX 565, 7550 AN Hengelo pemmerik@solair1.inter.NL.net The Netherlands Phone: +31 74 2558 688 --
From: mtrombin@ix.netcom.com (Mark Trombino) Newsgroups: comp.sys.next.programmer Subject: Re: NSWindow hangs in new thread Date: 8 Sep 1997 07:17:04 GMT Organization: Egghead Billy, Inc. Message-ID: <5v08pg$t4v@sjx-ixn7.ix.netcom.com> References: <5uurvi$on5@dfw-ixnews12.ix.netcom.com> <34132b5f.3993642@news.wco.com> <5uvtp8$qbh@dfw-ixnews11.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii In <5uvtp8$qbh@dfw-ixnews11.ix.netcom.com> Mark Trombino wrote: > In <34132b5f.3993642@news.wco.com> Mike Paquette wrote: > > >Is there anything that would cause a window that was order to the front of > > >the screen in one thread and then ordered off screen in a separate thread > > It could be that the DPS context creating the window isn't the same as > > the one that's trying to terminate the window. I suggest that the > > easiest way to handle this might be to have the worker thread message > > back to your main event handling thread to close the window. You can > > use Distributed Objects to do this almost transparently. That was the trick... Using DO, I had the main thread remove the window instead of the "worker" one. Worked perfectly. So what have I learned? Only the thread that creates a window can remove it. Now there's somethin' ya don't learn everyday! Thanks again, Mike. -- |-------------------------------------------------------------------------| | Mark Trombino | J A M S o f t | | mtrombin@ix.netcom.com | Audio DSP Tools for Openstep & Rhapsody | |-------------------------------------------------------------------------|
From: "David \"Shoe Lace\" Pyke" <dpyke@deakin.edu.au> Newsgroups: relcom.comp.gis,rec.games.xtank.programmer,rec.games.programmer,rec.crafts.polymer-clay,fj.comp.image,de.comp.shareware,de.comp.os.os2.programmer,de.comp.os.ms-windows.programmer,de.alt.shareware,comp.windows.ms.programmer,comp.unix.sco.programmer,comp.unix.programmer,comp.sys.psion.programmer,comp.sys.next.programmer,comp.sys.newton.programmer,comp.os.os2.programmer,alt.comics.image Subject: Re: Converting TIF to JPEG Date: Mon, 08 Sep 1997 11:18:32 +1000 Organization: DEAKIN UNIVERSITY (Geelong) Vic, Aust. Message-ID: <34135268.29BF@deakin.edu.au> References: <NEWTNews.872476934.4923.infolink@infolink-2> <5u43uu$7s1$1@wagner.videotron.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > On 08/24/97, janetom@hotmail.com wrote: > > > > Hi all. > > > > > > I have more than 1,000 TIF images, and I want to convert them to > JPEG (in > > order to watch then at the IntraNet). > > > > > > How can I do it ? > > Is there is any program available for this batch converting ? > > > > Please Help... > > > > > > Jane Eden > > Email: janetom@hotmail.com > > > ========================================= L-view pro will also do it just fine. its at www.shareware.com David, The SSSSS L S S H H OOOO EEEEEE L AA CCCC EEEEEE S H H O O E L A A C C E SSSSS HHHHHH O O EEEEE L A A C EEEEE S H H O O E L AAAAAA C E S S H H O O E L A A C C E SSSSS H H OOOO EEEEEE LLLLLLL A A CCCC EEEEEE # ------------------------------------------------------------------------------ Check out my WWW HomePage at -> http://www.deakin.edu.au/~dpyke ------------------------------------------------------------------------------ PetErpiPerpIckeDapeCkofPickLedpEppeRshoWmanYpicKledPeppErsdIdpeTerpIperPicK
From: drifterusa@macconnectNOT.com (John Bauer) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sun, 7 Sep 1997 23:41:13 -0500 Organization: KAOS Message-ID: <19970907234113405260@accs-as15-dp07.dlls.grid.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> Christian Brunschen wrote: > In another post of mine I remarked that one should not read too harsh > 'rejection' into the words 'We have no plans to', which were what CFO > Anderson said about a few things -- such as, CHRP licensing, and > licensing MacOS 8. > > Interestingly, no less than 3 cloners have announced shipping boxes > with MacOS 8, or acquiring a license to do so -- UMAX, PowerTools and > .. Power Computing. Sort of corroborates my guess that 'We have no > plans to ...' meant, generally, 'no comment, but here's a phrase you > can print and feel like you have an answer'. Point taken as to not overreacting to statements that begin with "We have no plans to..." Plans change. However, note that Anderson's statement that Apple has no plans to license Mac OS 8 *for CHRP* is in no way refuted by later announcements from UMAX, PowerTools and Power Computing announcing that they would ship their products with Mac OS 8. <snip> > In _theory_, nothing prevents anyone from doing anything with any > OS. Yet often users find that certain tasks are more easily > accomplished in one environment, compared to others. You know, one of > the reasons why people, and companies, choose Macintoshen instead of > Wintel boxes in the first place. This is the point: *very few* companies chose Macintoshes despite the platform's superior ease of use and inherent long-term cost advantages. Apple's mounting credibility problems may make even fewer consider Rhapsody no matter how wonderful it is. As the history of Apple shows time and again, perception outweighs technological merits. -- John Bauer <remove NOT from email address above>
Subject: Re: Objective-C Fragile Base Class (long) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <872756559.2813@dejanews.com> <x7yb5mh60k.fsf@tom.ics.ele.tue.nl> <872781306.21356@dejanews.com> <jesjones-ya02408000R0209972131290001@news.halcyon.com> <5ul48j$3dj$1@owl.slip.net> <5ulae3$imb$1@gte1.gte.net> <340e6ca4.0@epflnews.epfl.ch> <5umnvq$ju9$1@gte2.gte.net> <34100283.0@epflnews.epfl.ch> <5uph79$pvu$1@gte1.gte.net> In-Reply-To: <5uph79$pvu$1@gte1.gte.net> From: marco@sente.ch (Marco Scheurer) Organization: Sen:te Message-ID: <3413c704.0@epflnews.epfl.ch> Date: 8 Sep 97 09:36:04 GMT On 09/05/97, G. Apple wrote: > [...] >> The use of + you described is bad practice. What do you get? It is >> just as much work to overload + as it is to define a new function, >> such as "translate", or maybe "tr" or "t", which are symbols I've >> found in geometry texts to denote a translation. You take risks >> because it is ambiguous: it could very well be understood has a form >> of union. You loose readability. >> >> I undertand that some people find overloading operators in general, >> and + in particular, at the slightest reason, cute, or cool. >> Unfortunately, these people often include programmers. >> > > Ok, so maybe I missed the target at whom I was aiming. My real > target was the chorus of objectors to any kind of operator > overloading or multiple inheritance. It seems to me that most of > these are in the Objective-C and Java camps, many of the latter > particularly who have jumped on Java because they have heard it was > easy and then believe "Now I are a programmer too, so don't confuse > me by changing the meaning of a plus sign". Apparently, that shoe > does not fit you, so I offer my apologies. (Like the cartoon says, > on the Internet no one knows that you're really a dog. :-) No harm done. I was objecting specifically to your claim that using + to translate rectangles was intuitive, and better than proposed alternatives, not to operator overloading in general (although this too is criticizable, see below). > > Although I agree with your comments about "slightest reason, cute, > or cool", I do not agree with that (properly used) operator > overloading make things "ambiguous" or "lose readibility". For > example, use of complex numbers and matrices in C++ can make an > engineering algorithm vastly more understandable than a bunch of > Fortran gobbledygook required to do the same thing. > [...] The key phrase here is "properly used". Using + to add vectors in C++ maybe acceptable, because it respects the semantics generally associated with + in math (a nearly universal convention, which should take precedence over your own) while using + to translate rectangles is not. Using + for THIS purpose IS ambiguous (and not proper use) because it changes the meaning of + to something arbitrary. If you do so, you could just as well use + to construct polygons (by adding points to a list). So yes: "don't confuse me by changing the meaning of a plus sign". I would agree with your use of overloading + and * for complex numbers and matrice operations. BUT there is still a problem of cohesion: for instance with vectors, if you use + for addition which symbols do use for the inner and outer products? Then, operator overloading as it is proposed in C++ is too limited to be really useful. C++ is not APL (at least something positive about C++ !?). I agree that proper use of operator overloading can increase readability, but this is mostly due to the infix notation: a + b is more natural than Plus (a, b), or a->Plus(b). I find Objective C's [a plus: b] quite readable. Now for translating rectangles, compare r + v with [r moveOriginTo:p] or [r translateBy:v]. -- Marco Scheurer (marco@sente.ch) Sen:te
From: d89cb@efd.lth.se (Christian Brunschen) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 09:39:41 GMT Organization: Lund Institute of Technology, Sweden Message-ID: <5v0h4t$1ir$1@news.lth.se> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> NNTP-Posting-User: d89cb In article <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com>, Daniel L. Taylor <LTaylor7@ix.netcom.com> wrote: >In article <5uur7i$9po$1@news.lth.se>, d89cb@efd.lth.se (Christian >Brunschen) wrote: > >> >> Well, whatever specs you have seen, probably define what _ine_ person, >> or company, saw as 'the NC' -- ie, a product they want to build and >> want to create some sort of hype for; but I can't accept that that >> definition would define what NC:s are in general. > >NC's have been generally proposed as low-cost diskless stations which >survive off of a server and lower the per-seat costs for corporate >computing by lowering the initial cost in addition to administration costs >(centralized setup, simplified Java-OS). Yes. But noone has really suggested that the _CPU_ be crippled as well. (and adding a local disk should usually be seen as adding a client-side 'persistent cache', so fits within the concept) > >I'm not denying that you can build a NC that has more power at the seat, >but the idea of NC's is to move away from client-heavy computing. No -- putting all the computing stuff on the server is not what NC:s are aimed at. NCs have their own CPUs specifically so that work is done where it needs to be done, but they get their software and administration info from a central server to keep those costs down. You're thinking about something more akin to 'X Terminals' (or 'Windows Terminals' like MicroSoft is suggesting ..) >The applications don't run or exist on the other platform (hell, the other >platform doesn't even yet exist) and you're assuming a) that they will be >easier to use and administer and b) that it would matter if they really >were. Mac OS is an order of magnitude easier to use and administer than >Wintel and already has a jumpstart on applications, yet which dominates >corporations? Know why? Basically for fears that Apple just managed to >reinforce. Under OPENSTEP, today, the usual 'office productivity applications' are available -- not Brand-Microsoft, but from other sources. The administration of an antire network of hundreds of machines running NeXTstep (an earlier version of what is now called OPENSTEP) has been demonstrated as being easily doable by a single person, centrally. And that's when the moniker 'NC' wasn't even a small buzzword. > >> >Straddle the line that close to a PC and you have the same costs, which >> >means your NC has *no* benefit *what-so-ever* over a standard Wintel PC. >> >> Oh yes -- the software: the User interface, the Administration >> software, the underlying OS .... > >*What* software??? Mac OS has 12,000 applications, but how many are >available for NC's? How many Java apps currently exist? OpenStep >> apps? Oh, quite a few :) There isn't the same overabundance -- ie, there are usually only one or two, maybe three, apps for any given task -- but in return, those applications are usually _very good_. >What percentage of the Wintel library is mirrored by the Mac OS library, >and what percentage by the Java library? The OpenStep library? What 'library' are you talking about -- the library of available software ? There may not be as many programs, applications, for MacOS or OPENSTEP as there are for Wintel, but these fewer applications manage to cover largely the same areas of work as on Wintel. > >Jobs is most likely going with an x86 NC. Any evidence for this besides rumors ... ? Any evidence at all for your repeated assertation that Jobs will direct his attentions at _one marketonly_ ? >What apps will it have? Will it >dual boot Windows? That's interesting...now it needs a local OS. If so, >why not just use Wintel PCs? Ease of use and costs alone, eh? > >If ease of use and lower costs aren't enough to put Mac OS, with 12,000 >applications, into corporations, what makes you think it's enough to put >NC's after Apple has dropped the desktop, PowerPC, and screwed more >partners? Is there any proof that Apple 'has dropped the desktop' ? Is there any proof that Apple 'has dropped the PowerPC' ? No on both those counts. Yes, Steve Jobs has said that Microsoft has won the war. But that just means that Apple won't try to 'topple Microsoft', and does not mean that Apple has abandoned the desktop. As for abandoning the PowerPC, well, Apple probably wants to continue using PowerPC as their 'major' 'native' platform. But, by having a solution that is not _bound_ to one CPU family, they do not need to give in to any hardball play from Motorola, for instance -- as has been _rumored_, but not proven. > >Under different conditions I would agree with you. But Apple has blown >it's feet and lower legs out from under it with the recent move to close >and abandon the Mac market. That puts any Apple NC at a major >disadvantage. > >> (Just to set one thing straight: I am not a 'NC and Java fan' in the >> sense you seem to paint them here -- I 'envision' an 'NC' in the sense >> of a rather heavily loaded, hardware-wise, Rhapsody machine (be it >> Intel-, PPC-, Alpha-, SPARC-, MIPS-, PA-RISC- or whatever-based), >> which is centrally administered, gets its software 'pushed' to ir from >> a central server, and which mounts users' home directories remote, but >> which has enough secondary storage that jobs that need to do lots of >> stuff with lots of data on dis, can do that locally. Nothing that >> OPENSTEP machines don't offer you today, actually -- but by using the >> 'NC' moniker, you get people to actually _do_ that, instead of setting >> each machine up as a 'Personal Workstation' and just connecting them >> loosely to a network.) > >This is *not* the traditionally defined NC. You are describing a "client >heavy" PC, basically. Is there a 'spec' for this ? as you seem to be looking towards your specs for NCs so often. > >Which might sell well except you're missing the point's that: >* Apple just screwed more people in another round of "we promised this but >will not deliver", shaking any faith anybody has in them even more. Wrong - plenty of people still have lots of faith in Apple >* We now do not know what Apple's NC will look like. Client heavy super >machine or diskless station running what OS and capable of what? There is indeed no indication what an Apple NC will look like -- nor that there will be only one. Nor that NC's will replace Apple's current product line. >* Apple appears to be ditching both PowerPC and the desktop, losing four >major advantages (Mac OS customer base, Mac OS compatibility - which won't >mean anything without a customer base and growing software library, >superior price/performance, and a scalable strategy that lets them service >a wide range of needs). Assertations based on rumors, but not on provable fact. > >> Well, Apple's development of Rhapsody looks arather like they will be >> putting their money not with the 'Java-based, underpowered NC's you >> seem to think of, but rather in the area of 'centrally administered, >> powerfol workstation-class machines' that I envision :) OPENSTEP >> itself is somewhat resource-hungry, so by necessity a Rhapsody-based >> 'NC' would have to have sufficient resources to run both Rhapsody >> itself, and whatever applications the users want to run on top of >> this. > >Amelio's plan has been scratched in favor of pure NC's. And we don't >actually know that they will run Rhapsody. They may, rather, run Newton >OS. I'm not talking about 'Amelio's Plan' , I am talking about Apple's _actions_. > >My point isn't that NC's are bad, it's that ditching the Mac market and >going for pure NC's will kill Apple. As Apple surely knows -- and therefore they won't 'go for pure NCs'! In case you hadn't noticed, Rhapsody is coming nicely along, expected to be released into DR1 before the end of the month. Rhapsody is _not_ the stuff that 'pure NCs' are made of. >> In another post of mine I remarked that one should not read too harsh >> 'rejection' into the words 'We have no plans to', which were what CFO >> Anderson said about a few things -- such as, CHRP licensing, and >> licensing MacOS 8. >> >> Interestingly, no less than 3 cloners have announced shipping boxes >> with MacOS 8, or acquiring a license to do so -- UMAX, PowerTools and >> .. Power Computing. Sort of corroborates my guess that 'We have no >> plans to ...' meant, generally, 'no comment, but here's a phrase you >> can print and feel like you have an answer'. > >Power Computing is being bought by Apple and there is no point in denying >them OS 8. > >Power Tools may just be shipping the box...we don't know yet. > >UMAX just has OS 8, no CHRP. > >Motorola is screwed. > >There are no other cloners. Counterexample: PowerTools. Who are, as we speak, _shipping_ MacOS clones with MacOS 8, with Apple's blessing. (I mentioned them above, too.) > >This is *not* an open market and I think the Mac base has every right to >read it harshly. > >> >No, I grasp it perfectly. What you fail to grasp is that there is nothing >> >preventing MS and Intel, and 3rd party developers, from making Wintel PC's >> >about as easy to administer. If that's even a large factor. Look at how >> >well this argument worked for Mac OS in corporations. >> >> Oh, there is -- Microsoft Windows :) > >LOL! OK, so you made me eat that last statement. Well, it's party facetious, but party serious as well -- Microsoft Windows' (95, NT, etc) innards appear to be such that easy remote, centralized administration is not easily added. Plus of course, the GUI is not quite as consistent as, say OPENSTEP's or MacOS's, adding another level of 'complexity' (in a loose sense of the word) > >> In _theory_, nothing prevents anyone from doing anything with any >> OS. Yet often users find that certain tasks are more easily >> accomplished in one environment, compared to others. You know, one of >> the reasons why people, and companies, choose Macintoshen instead of >> Wintel boxes in the first place. > >But these 60 million users of 28 million machines may not stick around >long enough to even see if Apple's new machines can fit their needs. You basically claim that Apple's recent actions have pissed off _all_ of Apple's customers, so that they won't buy any new Apple products. This is simply not true: I work in a place where Macs abound, and noone here much cares about the recent actions; they will keep using, and buying, new Apple products. > >> Even if we compare today's Windows offerings (95 and NT) against >> today's 'Rhapsody-equivalent' (OPENSTEP 4.2), > >Yet OpenStep was all but dead when Apple bought it and Windows >> domiantes. Windows dominates, yes -- in some markets. But OPENSTEP was far from 'all but dead'. It was moving along nicely, in fact. > >Don't you get why I'm worried yet? Apple is making the same lame-brained >business decisions all over again. Those decisions would kill even the >computer off the Enterprise-E. Rhapsody and NC's don't stand a chance with >business moves like these.... That largely remains to be seen - and, of course, opinions differ. Many people actually think that Apple's recent decisions are _good_ ones. >> >> Since I firmly beleive that Apple's solution will be better than that >> offered by Wintel, I would purchase an 'Apple branded NC'. And I >> further beleive that enough people and companies besides me will come >> to similar conclusions. Sufficiently many, even, to keep Apple a >> viable company, and Rhapsody a viable platform. > >Will they? You're speaking in a newsgroup full of Mac faithful, who know >how much better Apple's technology is, yet they are ready to leave due to >business screw-ups. 'They' aren't --- some of the more outspoken angry ones are. Not, by far, all of them. Also remember, that those who are really 'fanatic' about something (and claim to be 'faithful') often are indeed more fanatic than faithful. (I have a historic example about this, but then I would have lost the argument, according to Godwin's rule.) > >> 'Even' you ? Considering there are still lots of people who do trust >> Apple enough to want to buy their systems (hardware and software), >> your threshold may not have been as high as you thought. > >We will not know how "many" until the complete fallout from killing clones >is felt, during the fall sales season. Clones are not _dead_, as UMAX and PowerTools clearly show. > >My guess is that there will not be very "many" left. My guess is that there will. >> >> I disagree on several points. Not everyone distrusts Apple the way you >> claim they do -- opinions differ on the issues; some people, including >> well-renowned indostry analists, consider Apple's recent moves to be >> good moves -- for Apple. > >I see a lot of customers, Mac faithful, walking away now. I see a lot of customers, Mac faithful, _remaining_ Mac faithful. > >Analysts can say what they will...they also said Apple should have died 10 >years ago...if they are now positive about Apple, it really could be the >end! *chuckle* you may have a point :) > >> >And sorry to burst your bubble, but who is going to buy Apple's version >> >over Microsoft's? >> >> The same people who currently buy Apple's products now over Microsofts >> -- those who buy what they need rather than what they are told they >> need ? :) > >The same ones voting against Apple in online polls and posting messages >like "F*** you Jobs! I'm buying Wintel!".... Well, _let them_ -- they're the same rabid people who have given Macintosh advocacy a bad name. Now the Wintel crowd will have to bear with them. Happy Happy Joy Joy :) But the core of those who buy Macintosh because they need them for their work, rather than because it's a suitable crusade, will remain. Because the Quality of the Mac platform has recently gone _up_ rather a lot -- MacOS 8 is the most stable yet -- and Rhapsody is on the horizon, and far more visible than Copland ever was. Religion is one thing -- but it is the technological advantage that will win people over to Apple. > >-- >Daniel L. Taylor >LTaylor7@ix.netcom.com >================================================================= >The Heritage Foundation /// Town Hall >http://www.heritage.org/heritage/ /// http://www.townhall.com/ >================================================================= // Christian Brunschen
From: hamm@spamsquish.ix.netcom.com (John Hammett) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: Mon, 08 Sep 1997 11:24:28 GMT Organization: Face the Muzak! Message-ID: <3414df91.1286955323@news.alt.net> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> <mschemen-0409972046570001@ts005d05.ftw-tx.concentric.net> <3411AFD8.8EF55F7D@spamtoNull.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Jonathan, I just want to say I've been reading your several dozen daily posts here, and I don't know where you get the energy. You've been knocking down flies right and left, with your logic and facts. Thanks a lot; I don't have to work so hard these days now that you're here doing battle with the loonies out there. Keep up the good work -- and don't forget to take a vacation now and then! -- John Hammet Jonathan Harker <Limeyville@spamtoNull.com> wrote, >And if what you say is true, then every person who bought 0S8 off the >shelf was likewise subsidized by Apple to the tune of "several hundred >dollars" since, like clone companies, they bought no new Apple hardware >to go along with it. > >Secondly, your viewpoint is incorrect because it assumes Apple received >no payment for the OS from clone companies. In fact, not only did the >clone companies pay $50 for OS 7.6, they also paid an average of $150 >per system to Apple for hardware licensing fees. That means, according >to you, that Apple "subsidized" off-the-shelf OS8 buyers for even *more >money* than it "subsidized" clone companies. > >Of course, this is all balderdash because Apple subsidized nothing to >anyone. Apple was paid by everyone, not the other way around. Is it >your suggestion that, in the absence of cloners, Apple will now SLASH >R&D expenditures to eliminate the "subsidies" it had been "paying" to >cloners? Let's see, they ought to be able to cut back R&D expenses by >about half now that the cloners are gone. Yea, right. > > >> >> Of course they are going to compare it to the legal contract and not an >> agreement that was never finalized by a CEO that is no longer with Apple. >> That does not make it a lie. > > >No, but it does make it a clear breach of contract. The pre-agreements >were written, signed documents that were *binding* on not just the >cloners, but on Apple as well. Why do you think Apple paid off Power to >the tune of a whopping $100M? Apple may not be so lucky with Motorola. > > > >> >> According to the expert Harry Norr, Apple was already getting 100 to 280 >> per box when you concider software and hardware fees. So a jump to 100 to >> 500 does not seem like a jump at all. > > >Right. Sure. Sell one million machines @ $280 and pay Apple $280M. Or >pay $500M on the same number of sales. A paltry difference of only 220 >MILLION DOLLARS. Yea, buddy--no jump at all. > > >> >> I think Gil got fired because he agreed to terms that were not much better >> then the original terms. At least not enougth better to satisfy the board. > >Don't forget the "board" got "fired" too--or most of them.
From: alex@witty.com (Alex Curylo) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Mon, 08 Sep 1997 08:17:39 -0700 Organization: Bavarian Illuminati Message-ID: <alex-0809970817390001@p1-40.van.tvs.net> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> <5utdrs$41l$1@attila.apana.org.au> <19970907224459112636@slip22126.rmii.com> <5v04b4$h0a$1@attila.apana.org.au> In article <5v04b4$h0a$1@attila.apana.org.au>, mjm@attila.apana.org.au (Matt McLeod) wrote: > Then perhaps Apple (and other commercial software developers) should learn > from the freeware experience. Um, excuse me here, but the key word is "free". The "freeware experience" is a HOBBY, by which I mean "something done for personal satisfaction which is not necessary to support one's lifestyle". Observing that lots of hobbyist programmers produce a great amount of code is about as insightful as observing that lots of hobbyist birdwatchers produce a great amount of bird sightings. If you're trying to develop software in order to feed your family and the families of the others in your firm, like myself, instead of for your own personal satisfaction, like the collective Linux community, then it seems to me there's not too much of value to be learned from this other than the truism that many hands make light work. But if you have any specific suggestions on how to harness hobbyists's energies like Linus in support of a product that people's families ARE depending on revenue from to keep food on the table and a roof over their heads, please share ;) ---- Alex Curylo -- alex@witty.com
From: alex@witty.com (Alex Curylo) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Mon, 08 Sep 1997 08:24:16 -0700 Organization: Bavarian Illuminati Message-ID: <alex-0809970824160001@p1-40.van.tvs.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <341074F1.8F26E657@spamtoNull.com> <5urae7$f7c$1@attila.apana.org.au> <gesNOSPAM-0709971424390001@dialup45.sisna.com> <5v04hq$h1p$1@attila.apana.org.au> In article <5v04hq$h1p$1@attila.apana.org.au>, mjm@attila.apana.org.au (Matt McLeod) wrote: (And no, I don't beleive that it's this "NC" > thing - at least, not the diskless, Java-based, thing). However, I wouldn't take any bets that it's not the swapdisk, xxxStep-based, thing. Indeed, it seems from random c.s.next.* postings that lots of people DO run their xxxStep networks in pretty much the way NC networks are envisioned to work -- applications live on the server and whatever random machine you sit down and log in at, you get access to your home desktop as if you were there. That sounds like the essence of an NC network to me, and pretty darn appealing to boot for anyone who lives in an environment with more than one machine around. ---- Alex Curylo -- alex@witty.com
Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc From: nagle@netcom.com (John Nagle) Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Message-ID: <nagleEG77Bp.Kor@netcom.com> Organization: Netcom On-Line Services References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <5up1dc$ndg$1@news.lth.se> <341074F1.8F26E657@spamtoNull.com> <5urae7$f7c$1@attila.apana.org.au> <gesNOSPAM-0709971424390001@dialup45.sisna.com> Date: Mon, 8 Sep 1997 16:33:25 GMT Sender: nagle@netcom6.netcom.com gesNOSPAM@sisna.com (ges) writes: >All this rationalizing doesn't reassure me. I think Apple is committing >suicide, or rather Jobs is murdering it. Apple should go out of business while they still have a positive net worth. They should just say "we lost - it's over", instead of this death march to bankruptcy. The stockholders would still get something. As it is, they may lose everything. John Nagle
From: Philippe Converset <plc@ina.fr> Newsgroups: comp.sys.next.programmer Subject: OpenStep Linking error on bundle Date: Mon, 08 Sep 1997 18:44:00 +0200 Organization: INA Message-ID: <34142B50.41C6@ina.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Here is a very simple example which produces a link error in NT OpenStep Enterprise 4.2 : An application with a single class "ClassA" an a loadable bundle with a single class "ClassB". ClassB is using ClassA (sending a message in one of its methods). The result of compilation is : ClassB.o : error LNK2001 unresolved external symbol .objc_class_name_ClassA It works fine under Mach but not under NT. Did I forget a compiler or linker flag? Is NT not supporting loadable bundles? Is it a Project Builder bug? Any idea?????
From: Eren_Kotan@apple.com (Eren Kotan) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 17:12:28 GMT Organization: NeXT/Apple European Professional Services Message-ID: <5v1bls$tri$1@news2.apple.com> References: <5urae7$f7c$1@attila.apana.org.au> <B0372D47-F4AD@206.165.43.161> Cc: english@primenet.com In <B0372D47-F4AD@206.165.43.161> "Lawson English" wrote: > Nope. Ford Motor, and various other embedded market companies provide far > greater markets for PPC's than Apple's measely little 4 million/year. Indeed, and before anybody else writes in to ask, let me just say that we have no plans at the moment to port Rhapsody to Fords. ;-} Regards, Eren -- Eren Kotan - E-mail: Eren_Kotan@apple.com The best friend money can buy. Can you imagine a world with no hypothetical questions? Apple Enterprise Professional Services - formerly NeXT Software
From: Rich Schroedel <richs@win.bright.net> Newsgroups: comp.sys.next.programmer Subject: Exception processing in Objective C? Date: Tue, 02 Sep 1997 13:42:00 -0600 Organization: BrightNet Wisconsin Message-ID: <340C6C07.5727@win.bright.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I'm in the process of rewriting some C++ code so that it can eventually be converted to Objective C. I'm stumped at how to handle exception processing. How do Objective C programmers handle the sort of situations that C++ programmers handle with exception processing? Any help out there? -- Rich Schroedel "There is only one success... Ondossagon Software to live your life in your own way" richs@win.bright.net Christopher Marlowe
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Linking error on bundle Date: 8 Sep 1997 17:57:38 GMT Organization: Anderson Financial Systems Inc. Message-ID: <5v1eai$lni@shelob.afs.com> References: <34142B50.41C6@ina.fr> Philippe Converset <plc@ina.fr> writes > Here is a very simple example which produces a link error in NT OpenStep > Enterprise 4.2 : > > An application with a single class "ClassA" an a loadable bundle with a > single class "ClassB". ClassB is using ClassA (sending a message in one > of its methods). > > The result of compilation is : > ClassB.o : error LNK2001 unresolved external symbol > .objc_class_name_ClassA > > It works fine under Mach but not under NT. > Did I forget a compiler or linker flag? > Is NT not supporting loadable bundles? > Is it a Project Builder bug? The problem is that the NT linker does not observe the -ObjC or -include_all linker symbols. Therefore, it only brings in classes to which there are explicit references elsewhere in your source code. (This bit me a few times with classes that only exist in NIB files.) The solution is to refer to the missing class explicitly in some piece of code that you know will be included. I usually put a harmless [ClassA class] statement in an +initialize block. That simple reference will force the linker to resolve the ClassA symbol. -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Exception processing in Objective C? Date: 8 Sep 1997 18:09:11 GMT Organization: Anderson Financial Systems Inc. Message-ID: <5v1f07$loh@shelob.afs.com> References: <340C6C07.5727@win.bright.net> Rich Schroedel <richs@win.bright.net> writes > I'm in the process of rewriting some C++ code so that it can eventually > be converted to Objective C. I'm stumped at how to handle exception > processing. How do Objective C programmers handle the sort of situations > that C++ programmers handle with exception processing? Any help out > there? In OPENSTEP and Rhapsody, NSException is a class unto itself. I don't do C++, but it is remarkably similar to my experience with structured expections in Delphi (where Exceptions are also wrapped by a class). -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Mon, 08 Sep 1997 11:50:03 -0700 Organization: Netcom Message-ID: <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> <5v0h4t$1ir$1@news.lth.se> In article <5v0h4t$1ir$1@news.lth.se>, d89cb@efd.lth.se (Christian Brunschen) wrote: > >NC's have been generally proposed as low-cost diskless stations which > >survive off of a server and lower the per-seat costs for corporate > >computing by lowering the initial cost in addition to administration costs > >(centralized setup, simplified Java-OS). > > Yes. But noone has really suggested that the _CPU_ be crippled as > well. (and adding a local disk should usually be seen as adding a > client-side 'persistent cache', so fits within the concept) The architecture is "crippled" from the outset. No expansion, little RAM, no disk/small disk, all to keep costs down. Things aren't going to move that quickly over Ethernet, even the 100Mb/s variety. Now, for some needs this is more than sufficient. But it's not something you bet all of Apple on, if that is indeed what Jobs is doing. In my last post I kept saying that is what's happening, and I shouldn't of because there's been no official announcement. I apologize for that. Indeed, we don't know what's happening at all. Which means I can't see anybody recommending Mac OS right now, because without a clear indication of strategy from Apple it's a very risky proposition. Which is yet another example of just how bad Apple's business moves have been since Steve has taken over. But, that's besides the point...or rather, maybe it *is* the point. > >I'm not denying that you can build a NC that has more power at the seat, > >but the idea of NC's is to move away from client-heavy computing. > > No -- putting all the computing stuff on the server is not what NC:s > are aimed at. NCs have their own CPUs specifically so that work is > done where it needs to be done, but they get their software and > administration info from a central server to keep those costs > down. Yes, again, the idea is to move away from "client-heavy" computing. > >The applications don't run or exist on the other platform (hell, the other > >platform doesn't even yet exist) and you're assuming a) that they will be > >easier to use and administer and b) that it would matter if they really > >were. Mac OS is an order of magnitude easier to use and administer than > >Wintel and already has a jumpstart on applications, yet which dominates > >corporations? Know why? Basically for fears that Apple just managed to > >reinforce. > > Under OPENSTEP, today, the usual 'office productivity applications' > are available -- not Brand-Microsoft, but from other > sources. The administration of an antire network of hundreds of > machines running NeXTstep (an earlier version of what is now called > OPENSTEP) has been demonstrated as being easily doable by a single > person, centrally. And that's when the moniker 'NC' wasn't even a > small buzzword. OK, you've missed the point. "Mac OS is an order of magnitude easier to use and administer than Wintel and already has a jumpstart on applications, yet which dominates corporations? Know why? Basically for fears that Apple just managed to reinforce." Again, the best OS in the world will die if Apple continues to do business the way it is doing business right now. > >*What* software??? Mac OS has 12,000 applications, but how many are > >available for NC's? How many Java apps currently exist? OpenStep > >> apps? > > Oh, quite a few :) There isn't the same overabundance -- ie, there are > usually only one or two, maybe three, apps for any given task -- but > in return, those applications are usually _very good_. Or non-existent. Currently there are several areas where Mac OS will not do as a solution simply due to lack of software. If *Mac OS* has this problem, with 12,000 apps, then OpenStep/Java are in *big* trouble. Rhapsody was to leverage the Mac OS library as a starting point with the Blue Box. Now, who knows what will happen.... > >Jobs is most likely going with an x86 NC. > > Any evidence for this besides rumors ... ? > > Any evidence at all for your repeated assertation that Jobs will > direct his attentions at _one marketonly_ ? I consider the end of Mac cloning to be evidence of that, as well as the way Apple has screwed its PPC partners. If not evidence of his intentions, this may well be an indication of what will happen any way. > >What apps will it have? Will it > >dual boot Windows? That's interesting...now it needs a local OS. If so, > >why not just use Wintel PCs? Ease of use and costs alone, eh? > > > >If ease of use and lower costs aren't enough to put Mac OS, with 12,000 > >applications, into corporations, what makes you think it's enough to put > >NC's after Apple has dropped the desktop, PowerPC, and screwed more > >partners? > > Is there any proof that Apple 'has dropped the desktop' ? Apple has recently killed the Mac clone market, closing the market to the dismay and anger of a huge portion of the Macintosh base. Apple has also abandoned CHRP, considered critical to Macintosh success on the desktop. In addition, Steve himself declared the "war over" at MacWorld Boston. > Is there any proof that Apple 'has dropped the PowerPC' ? This is not Apple's choice, but IBM and Motorola's. Based on how they've been screwed by Apple, it wouldn't be a surprise. > >This is *not* the traditionally defined NC. You are describing a "client > >heavy" PC, basically. > > Is there a 'spec' for this ? as you seem to be looking towards your > specs for NCs so often. Look, NC's have not been traditionally described or defined as you describe them. Taking a Power Mac and administering its software from a server *does not* make it a NC by most people's standards. > >Which might sell well except you're missing the point's that: > >* Apple just screwed more people in another round of "we promised this but > >will not deliver", shaking any faith anybody has in them even more. > > Wrong - plenty of people still have lots of faith in Apple Whatever. One look at this newsgroup should be enough to send chills up Apple's spine.... > >* Apple appears to be ditching both PowerPC and the desktop, losing four > >major advantages (Mac OS customer base, Mac OS compatibility - which won't > >mean anything without a customer base and growing software library, > >superior price/performance, and a scalable strategy that lets them service > >a wide range of needs). > > Assertations based on rumors, but not on provable fact. You know what? I can no longer live by ignoring the clues and just hoping it will all turn out right. Do you really want me to honestly recommend a Mac OS system to *anybody* at this point? Apple has smashed their original strategy under Amelio, one I put great faith in. And with no clear strategy to replace it, why on Earth should I stick by? Why should anyone? Yeah, my theory of what's happening is based on rumor at this point. But in the absence of a clear strategy that the customer base can evaluate and base purchasing decisions on, that's all we've got. Given that, you're in no better position than I am. You *cannot* prove that Apple has a solid strategy, or even *begin* to evaluate if they will succeed or not. And that makes Wintel look *very* attractive.... > >Amelio's plan has been scratched in favor of pure NC's. And we don't > >actually know that they will run Rhapsody. They may, rather, run Newton > >OS. > > I'm not talking about 'Amelio's Plan' , I am talking about Apple's > _actions_. Which have been exceedingly poor and confusing at this point. > >My point isn't that NC's are bad, it's that ditching the Mac market and > >going for pure NC's will kill Apple. > > As Apple surely knows -- and therefore they won't 'go for pure NCs'! Do they know this? Give me a press statement, or a paper articulating their strategy. Don't have one? Hmmm.... Well, I admire you for not giving in to rumors as I have (I admit this). But you know what? At least the rumors keep me interested and create a period of pause before I abandon the platform. If I were to make a cold, purely logical, based 100% on fact purchasing decision at this moment, I would be forced to default to Wintel in the absence of any stated plan on Apple's part to survive. In other words, Apple isn't even on the playing field right now. Is that preferable? > >Power Computing is being bought by Apple and there is no point in denying > >them OS 8. > > > >Power Tools may just be shipping the box...we don't know yet. > > > >UMAX just has OS 8, no CHRP. > > > >Motorola is screwed. > > > >There are no other cloners. > > Counterexample: PowerTools. Who are, as we speak, _shipping_ MacOS > clones with MacOS 8, with Apple's blessing. (I mentioned them above, > too.) Oh well, forgive me, *that's* an open market... :-/ > >But these 60 million users of 28 million machines may not stick around > >long enough to even see if Apple's new machines can fit their needs. > > You basically claim that Apple's recent actions have pissed off _all_ > of Apple's customers, so that they won't buy any new Apple > products. Not all. Most. > >Don't you get why I'm worried yet? Apple is making the same lame-brained > >business decisions all over again. Those decisions would kill even the > >computer off the Enterprise-E. Rhapsody and NC's don't stand a chance with > >business moves like these.... > > That largely remains to be seen - and, of course, opinions > differ. Many people actually think that Apple's recent decisions are > _good_ ones. Who? And no, Larry Ellison and Steve Jobs do not count.... > >We will not know how "many" until the complete fallout from killing clones > >is felt, during the fall sales season. > > Clones are not _dead_, as UMAX and PowerTools clearly show. CHRP is dead. Clones are effectively dead. > >> The same people who currently buy Apple's products now over Microsofts > >> -- those who buy what they need rather than what they are told they > >> need ? :) > > > >The same ones voting against Apple in online polls and posting messages > >like "F*** you Jobs! I'm buying Wintel!".... > > Well, _let them_ -- they're the same rabid people who have given > Macintosh advocacy a bad name. Now the Wintel crowd will have to bear > with them. Happy Happy Joy Joy :) Are you happy I'm considering a move to Wintel? > But the core of those who buy Macintosh because they need them for > their work, rather than because it's a suitable crusade, will remain. I need a viable computer platform for my work. What has Apple said or done that would convince me they will be viable in the future? You want me to stick to facts, so let's go with the facts: *Fact: Apple has severely curbed its original clone plans, and this will most likely result in higher prices and lower performance. This leaves Macintosh at a new disadvantage to Wintel. * Fact: Amelio's plan, which I believed to be solid, has now been tossed aside. * Fact: Apple has not presented, in any form, a strategy to replace Amelio's. Barring rumors, I guess I'm to act as if they have none. Hmmm.... > Religion is one thing -- but it is the technological advantage that > will win people over to Apple. Gee, where have I heard that before? Lisa, Amiga, Mac OS, NextStep.... You know what? It takes more than technological advantage. It takes solid business decisions and marketing. Under Amelio, I was seeing that start to happen. But in a very short time, Jobs as smashed all of that and gone back to horrible business decisions and nightmarish PR. That's fact. Not rumor, not my speculation (OK, so I've done too much of that)...just plain fact. Given those facts, I'm closer than ever to going Wintel. -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: shoppa@alph02.triumf.ca (Tim Shoppa) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 7 Sep 1997 16:06:43 GMT Organization: TRIUMF, Canada's National Meson Facility Message-ID: <5uujej$2to$1@nntp.ucs.ubc.ca> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <01bcba15$447c2f00$48e04382@broennic.online.no> <acurylo-0509971010020001@p2-29.van.tvs.net> <5ur2uv$r3s@crcnis3.unl.edu> Delivery-Date: Sun, 07 Sep 1997 09:12:30 -0700 Return-Path: news@nntp.ucs.ubc.ca for <gherbert@crl.com>; Sun, 7 Sep 1997 09:02:48 -0700 (PDT) env-from (news@nntp.ucs.ubc.ca) id QQdfzk11041; Sun, 7 Sep 1997 12:07:04 -0400 (EDT) To: sci-space-tech@moderators.uu.net In article <5ur2uv$r3s@crcnis3.unl.edu>, Josh Hesse <00093182@bigred.unl.edu> wrote: >Alex Curylo (acurylo@inmediapresents.nospam.com) wrote: >: >: The PowerPC family dominates a large number of embedded systems markets. >: Besides cars, there's a PPC in pretty much every piece of medical imaging >: hardware, for instance. Maybe if you restrict yourself only to new imaging hardware. There's a huge installed base of instrumentes based on PDP-11's out there... >: Not to mention the Mars Rover. (PPC -- 100% of the >: Martian computing installed base! There's an ad for you huh?) >: > >Not to rain on your extraterrestial parade or anything, but >doesn't the Mars Rover use a version of the Intel 8085? Yes, a 80C85 is in the Rover. The lander has a IBM RAD6000, which is vaguely related to the PowerPC (but it is *not* a PowerPC). Do we really want these sandbox-type "my CPU is better than your CPU" arguments spilling over into sci.space.tech? Tim. (shoppa@triumf.ca) [Mod note: The topic of the CPUs used in spacecraft is fine here; I have itchy trigger finger if this degenerates into other topics. You have been warned. -gwh]
From: madler@alumni.caltech.edu (Mark Adler) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 7 Sep 1997 18:26:19 GMT Organization: Caltech Alumni Association Message-ID: <5uurkb$d74@gap.cco.caltech.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <01bcba15$447c2f00$48e04382@broennic.online.no> <acurylo-0509971010020001@p2-29.van.tvs.net> <5ur2uv$r3s@crcnis3.unl.edu> Delivery-Date: Sun, 07 Sep 1997 11:44:15 -0700 Return-Path: news@gap.cco.caltech.edu for <gherbert@crl.com>; Sun, 7 Sep 1997 11:21:54 -0700 (PDT) env-from (news@gap.cco.caltech.edu) id QQdfzt17538; Sun, 7 Sep 1997 14:26:39 -0400 (EDT) In article <5ur2uv$r3s@crcnis3.unl.edu>, Josh Hesse <00093182@bigred.unl.edu> wrote: >doesn't the Mars Rover use a version of the Intel 8085? Yep. A rad-hard 80C85. Pathfinder doesn't have a PPC either. mark
From: ibhan@digitas.harvard.edu (Ishir Bhan) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 18:07:47 GMT Organization: Harvard University, Cambridge, Massachusetts Message-ID: <5v1etj$84r$1@news.fas.harvard.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <LTaylor7-070997 LTaylor7@ix.netcom.com (Daniel L. Taylor) writes: >My point isn't that NC's are bad, it's that ditching the Mac market and >going for pure NC's will kill Apple. Why do you always equate these two? Why do you think that going for NC's will mean ditching the Macintosh market (unless you think that eventually switching people over to Rhapsody will be "ditching the Mac market"). Apple will almost sell both types of machines. -- Ishir Bhan ibhan@digitas.harvard.edu http://digitas.harvard.edu/~ibhan
From: ibhan@digitas.harvard.edu (Ishir Bhan) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 18:01:15 GMT Organization: Harvard University, Cambridge, Massachusetts Message-ID: <5v1ehb$7fk$1@news.fas.harvard.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> <LTaylor7-070997 LTaylor7@ix.netcom.com (Daniel L. Taylor) writes: >But *replacing* Amelio's original plan with NC's and abandoning the >desktop and Apple's user base in the process ("...milk the Macintosh for >all it's worth and move to the next great thing...") is damned foolishness >and will get Apple buried. Why do you think that Jobs plan is to make NC's instead of Rhapsody boxes? Let's face it: the MacOS is really at its limit. Rhapsody for PC's and either a modified Rhapsody for or Newton OS for NC's is probably where Apple is headed. What is your problem with this? -- Ishir Bhan ibhan@digitas.harvard.edu http://digitas.harvard.edu/~ibhan
From: "Ed Deans" <eadeans@ibm.net> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 6 Sep 1997 14:03:38 -0700 Organization: Road Runner Message-ID: <5usggh$r4@proxye1.san.rr.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.n et> <B03540AE-CE02A@206.165.44.43> <jrwst30-0509971132340001@jrwst30.j.resnet.pitt.edu> <see-below-0609970348030001@ip183.mountain-view2.ca.pub-ip.psi.net> <wjtaggart-0609970748590001@crazyhorse-19.fuse.net> William J Taggart wrote in article ... >In article ><see-below-0609970348030001@ip183.mountain-view2.ca.pub-ip.psi.net>, >see-below@not-my-address.com (Matthew Vaughan) wrote: > >> In article <jrwst30-0509971132340001@jrwst30.j.resnet.pitt.edu>, >> jrwst30@pitt.edu (James Wright) wrote: >> >> > secondly, IBM relies on the Power >> > Arhictecture for their workstation and supercomputing machines ; the next >> > generation of these machines will contain massively parralel >> > configurations of top of the line PowerPc's. >> >> But they already make seperate chips for their workstations and servers. >> They don't necessarily use the same PPC chips as PowerMacs. >> > >I just finished an article about the new-and-improved IBM RS/6000. The >article stated that the RS/6000 used 604e microprocessors (not clear about >whether the 604e usage was part of the "new-and-improved"). How are these >different from the 604e cpus used in PPCs? These are the same 604e's just running which a much more efficient motherboard and i/o design than Apple's PowerMacs. The new RS/6000's (as best as I recall) are indeed new to the RS/6000 line. IBM has shown that the older PPC processors have lots of milage in them if you don't needlessly burden the processor with a poor i/o design. NeXT used to understand that *throughput* was the key--when's Apple going to realize this? Between Be's whiz-bang demos on modest Apple PPC hardware and IBM's whiz-bang capabilities with superior designs and the same CPU's it's no wonder that circles are being run around Apple. It's pittiful. --Ed.
From: Jonathan Harker <Limeyville@spamtoNull.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.misc,comp.sys.mac.system Subject: Re: Steve Jobs' 'BIG LIE" (Was Re: Apple's cloning plan fatally flawed. Date: Mon, 08 Sep 1997 17:48:03 -0400 Organization: Netcom Message-ID: <34147293.13EF191F@spamtoNull.com> References: <19970903204855473347@lc019.zianet.com> <B0338518-199D30@206.165.44.50> <mschemen-0409972046570001@ts005d05.ftw-tx.concentric.net> <3411AFD8.8EF55F7D@spamtoNull.com> <3414df91.1286955323@news.alt.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit John Hammett wrote: > > Jonathan, I just want to say I've been reading your several dozen daily posts > here, and I don't know where you get the energy. You've been knocking down flies > right and left, with your logic and facts. Thanks a lot; I don't have to work so > hard these days now that you're here doing battle with the loonies out there. > > Keep up the good work -- and don't forget to take a vacation now and then! I suppose I am fed up with hearing Jobs repeatedly and pathetically say, "Apple can't do this, and Apple can't do that" as a justification for his policies of retrenchment and retraction. All great companies and all truly great corporate pioneers say, instead, "Can do! Will do!" and then proceed along those lines with clear and definite aims. Jobs has been defeated so much in his own personal endeavors that now he's projecting that aura of defeatism onto Apple. It certainly shows. -- Jonathan Hollingsworth Harker New astronomical data shows the age of the universe as measured by the latest galactic red-shift data at between 8 and 12 billion years. However, hydrogen decay data proves indisputably that many stars are 15-20 billion years of age. Modern astronomical theory about the formation and age of the universe is so accurate that it tells us that many stars are older than the universe in which they exist. And you laughed when they said the world was flat?
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 14:51:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B039C692-2131D@206.165.44.76> References: <5v1bls$tri$1@news2.apple.com> nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.system, nntp://news.primenet.com/comp.sys.mac.misc MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Eren Kotan <Eren_Kotan@apple.com> said: > > In <B0372D47-F4AD@206.165.43.161> "Lawson English" wrote: > > > Nope. Ford Motor, and various other embedded market companies provide > far > > greater markets for PPC's than Apple's measely little 4 million/year. > > Indeed, and before anybody else writes in to ask, let me just say that we > have no plans at the moment to port Rhapsody to Fords. HOw's the dashboard version of PlainTalk coming along? ------------------------------------------------------------------------- Build a better mousetrap, yes, but then market it like crazy, or instead of catching mice you'll only collect dust. -David Yeargin -------------------------------------------------------------------------
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 14:55:01 -0700 Organization: Primenet Services for the Internet Message-ID: <B039C792-24F3D@206.165.44.76> References: <5v1ehb$7fk$1@news.fas.harvard.edu> nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.system, nntp://news.primenet.com/comp.sys.mac.misc MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Ishir Bhan <ibhan@digitas.harvard.edu> said: > Why do you think that Jobs plan is to make NC's instead of Rhapsody boxes? > Let's face it: the MacOS is really at its limit. Rhapsody for PC's and > either a modified Rhapsody for or Newton OS for NC's is probably where > Apple is headed. > > What is your problem with this? The margin for NC's is likely to be VERY slight. APple has had trouble in forcasting and meeting the demand for its far more profitable boxes (that's why the claim that Power was stealing from them makes NO sense -even with 50% margins, if you can't ship to your customers on time, a 50% margin on zero is still zero% profit) and it seems foolish to dedicate manufacturing capabilities that could go to the high-end to the lowest of the low-end. And if Apple expects clone-makers to line up for NC designs, who would now trust them to keep to agreements and step-forward? ------------------------------------------------------------------------- Build a better mousetrap, yes, but then market it like crazy, or instead of catching mice you'll only collect dust. -David Yeargin -------------------------------------------------------------------------
From: tim@jumpnet.com (Tim Olson) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 22:04:57 GMT Organization: At Right Angles Message-ID: <5v1sq9$dfb$1@news.jumpnet.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.ps i.net> <01bcba15$447c2f00$48e04382@broennic.online.no> <acurylo-0509971010020001@p2-29.van.tvs.net> <5ur2uv$r3s@crcnis3.unl.edu> <5uurkb$d74@gap.cco.caltech.edu> Delivery-Date: Mon, 08 Sep 1997 15:06:12 -0700 Return-Path: news@jump.net for <gherbert@crl.com>; Mon, 8 Sep 1997 15:00:32 -0700 (PDT) env-from (news@jump.net) for <sci-space-tech@moderators.uu.net>; Mon, 8 Sep 1997 15:05:09 -0700 (PDT) To: sci-space-tech@moderators.uu.net In article <5uurkb$d74@gap.cco.caltech.edu> madler@alumni.caltech.edu (Mark Adler) writes: > In article <5ur2uv$r3s@crcnis3.unl.edu>, > Josh Hesse <00093182@bigred.unl.edu> wrote: > >doesn't the Mars Rover use a version of the Intel 8085? > > Yep. A rad-hard 80C85. Pathfinder doesn't have a PPC either. Well, you are correct in the strictest sense of "PPC". But Pathfinder *does* have a rad-hard RS/6000 on board: [From IBM's web pages]: ---- The radiation-hardened RS/6000 forms the heart of the Attitude and Information Management subsystem, primarily built at JPL. "IBM developed the . . . space-compatible RS/6000," says Brian Muirhead, JPL's Pathfinder flight systems manager. Then Loral Federal Systems adapted it for Pathfinder. The RS/6000 will handle cruise altitude control and manage entry events, including airbag deployment. On Mars, it will run surface operations commanded from Earth. "We basically put all functions in one box, and the same team is responsible for all of it," says Rob Manning, Pathfinder's flight systems chief engineer. ---- -- Tim Olson
From: "Benjamin Smith" <benjs@ix.netcom.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 97 19:08:12 -0500 Organization: Netcom Message-ID: <B039FDA0-CBE63@206.214.136.134> References: <5v04hq$h1p$1@attila.apana.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit nntp://nntp.ix.netcom.com/comp.sys.next.advocacy, nntp://nntp.ix.netcom.com/comp.sys.mac.programmer.misc, nntp://nntp.ix.netcom.com/comp.sys.next.programmer, nntp://nntp.ix.netcom.com/comp.sys.mac.system, nntp://nntp.ix.netcom.com/comp.sys.mac.misc On Mon, Sep 8, 1997 1:04 AM, Matt McLeod <mailto:mjm@attila.apana.org.au> wrote: > >Despite the popular view (which I share) that management is made up of >idiots promoted beyond their competency, I don't really think that Apple is >completely stupid. Killing themselves off is not conducive to long-term >profitability. > >It looks to me like they've got some other "vision". Perhaps what they need >to do now is articulate it. (And no, I don't beleive that it's this "NC" >thing - at least, not the diskless, Java-based, thing). > The thing of it is is that the current people on Apple's board are proven. Many of them are from the industry. So, they have a plan that they think is sound. I guess many people believe that Jobs is calling all shots. I don't. And maybe they have another vision and this just is not the time to announce it. And there may be some very sound reasons behind it. I just wish that people would at least entertain this notion and not the familiar one that Jobs is trying to destroy Apple or he is destroying it through a series of wrong-headed moves. I think it is too early to tell. What's going on with the search for a CEO? Ben S.
From: mschemen@concentric.net (Mathew Schemenaur) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 8 Sep 1997 22:36:08 GMT Organization: Concentric Internet Services Message-ID: <mschemen-0709971745090001@ts004d19.ftw-tx.concentric.net> References: <5v1ehb$7fk$1@news.fas.harvard.edu> <B039C792-24F3D@206.165.44.76> In Motorolas press conference today, they did not Kill the PowerPC for the desktop. Instead they anounce that they would "broaden" they relationship with IBM to product PowerPC chips for the embedded market. This whole discussion based on junk. They guys are looking for the sky to fall and are willing to believe rumors even if they defy the laws of physics.
From: Tony Invencio <"tonyinv"@hotmail .dot com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Mon, 08 Sep 1997 20:53:34 -0400 Organization: Have None Message-ID: <5v26mf$fmu@dfw-ixnews11.ix.netcom.com> References: <5v04hq$h1p$1@attila.apana.org.au> <B039FDA0-CBE63@206.214.136.134> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Benjamin Smith wrote: > > On Mon, Sep 8, 1997 1:04 AM, Matt McLeod < > mailto:mjm@attila.apana.org.au> > wrote: > > > >Despite the popular view (which I share) that management is made up > of > >idiots promoted beyond their competency, I don't really think that > Apple > is > >completely stupid. Killing themselves off is not conducive to > long-term > >profitability. > > > >It looks to me like they've got some other "vision". Perhaps what > they > need > >to do now is articulate it. (And no, I don't beleive that it's this > "NC" > >thing - at least, not the diskless, Java-based, thing). > > > > The thing of it is is that the current people on Apple's board are > proven. > Many of them are from the industry. So, they have a plan that they > think is > sound. I guess many people believe that Jobs is calling all shots. I > don't. > > And maybe they have another vision and this just is not the time to > announce it. And there may be some very sound reasons behind it. > > I just wish that people would at least entertain this notion and not > the > familiar one that Jobs is trying to destroy Apple or he is destroying > it > through a series of wrong-headed moves. I think it is too early to > tell. > > What's going on with the search for a CEO? > > Ben S. Hey Ben would you want to be CEO around Jobs? Think about that before you answer. Unless you agree with all of Jobs's moves, I think any person would think twice before taking that Job. -- Tony Invencio Strange Solutions Software "The fault, dear Brutus, is not in our stars, But in ourselves, that we are underlings." Cassius
From: Jonathan Hendry <jon@subsequent.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 1997 01:16:02 GMT Organization: Netcom Sender: Jonathan Hendry <jon@aahz.ix.netcom.com> Message-ID: <5v280i$ks8@sjx-ixn5.ix.netcom.com> References: <5v1ehb$7fk$1@news.fas.harvard.edu> <B039C792-24F3D@206.165.44.76> Lawson English <english@primenet.com> wrote: > Ishir Bhan <ibhan@digitas.harvard.edu> said: > > Why do you think that Jobs plan is to make NC's instead of Rhapsody > boxes? > > Let's face it: the MacOS is really at its limit. Rhapsody for PC's and > > either a modified Rhapsody for or Newton OS for NC's is probably where > > Apple is headed. > > > > What is your problem with this? > The margin for NC's is likely to be VERY slight. APple has had trouble in > forcasting and meeting the demand for its far more profitable boxes (that's > why the claim that Power was stealing from them makes NO sense -even with > 50% margins, if you can't ship to your customers on time, a 50% margin on > zero is still zero% profit) and it seems foolish to dedicate manufacturing > capabilities that could go to the high-end to the lowest of the low-end. NC's would not be sold the same way as Macs, I'd bet. The NC would most likely be sold in quantity. A school or business would order, say a dozen at a time of a certain model. Apple would crank those out and ship them. That'd be fine for institutional buyers. If onesy-twosy sales are a factor, they'd likely also be handled direct from Apple. In either case, prediction isn't really an issue, apart from predicting if anyone will want a particular NC model at all. - Jon
From: "Lawson English" <english@primenet.com> Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: CPUs in Space (Was Re: No more PowerPCs... Date: 8 Sep 1997 14:48:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B039C5D9-1E794@206.165.44.76> References: <5uujej$2to$1@nntp.ucs.ubc.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Delivery-Date: Mon, 08 Sep 1997 15:37:38 -0700 Return-Path: root@nntp02.primenet.com for <gherbert@crl.com>; Mon, 8 Sep 1997 15:32:58 -0700 (PDT) env-from (root@nntp02.primenet.com) for <sci-space-tech@moderators.uu.net>; Mon, 8 Sep 1997 14:50:09 -0700 (MST) for <sci-space-tech@moderators.uu.net>; Mon, 8 Sep 1997 14:50:07 -0700 (MST) id OAA23340 for sci-space-tech@moderators.uu.net; Mon, 8 Sep 1997 14:50:05 -0700 (MST) To: sci-space-tech@moderators.uu.net nntp://news.primenet.com/comp.sys.mac.advocacy, nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.system, nntp://news.primenet.com/comp.sys.mac.misc Tim Shoppa <shoppa@alph02.triumf.ca> > Yes, a 80C85 is in the Rover. The lander has a IBM RAD6000, which > is vaguely related to the PowerPC (but it is *not* a PowerPC). > > Do we really want these sandbox-type "my CPU is better than your CPU" > arguments spilling over into sci.space.tech? > Just curious. The more modern CPUs with their denser transistor count, seem like they would be more vulnerable to things like errors due to cosmic rays. What kind of trade-off between vulnerability (if there is one), CPU-power, power consumption, size, etc., leads to the selection of a given CPU for such a mission? ------------------------------------------------------------------------- Build a better mousetrap, yes, but then market it like crazy, or instead of catching mice you'll only collect dust. -David Yeargin -------------------------------------------------------------------------
From: slee1@erols.REMOVE_ME.com (Steven Lee) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 00:12:33 -0400 Organization: Erol's Internet Services Message-ID: <slee1-0909970012330001@dam-as6s38.erols.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> <LTaylor7-070997 <5v1ehb$7fk$1@news.fas.harvard.edu> In article <5v1ehb$7fk$1@news.fas.harvard.edu>, ibhan@digitas.harvard.edu (Ishir Bhan) wrote: > Why do you think that Jobs plan is to make NC's instead of Rhapsody boxes? > Let's face it: the MacOS is really at its limit. Rhapsody for PC's and > either a modified Rhapsody for or Newton OS for NC's is probably where > Apple is headed. > > What is your problem with this? Quite a few. Here are the more important ones: 1) Macs have better hardware than Intel boxes. 2) PowerPC is way ahead of Pentium as a microprocessor. Just look at the benchmarks of Intel boxes vs Power Macs running Linux. Power Macs came out ahead. 3) People who own Mac hardware would end up with piles of very expensive door stops and paper weights when they move to the Intel platform. 4) What makes you think Steve Jobs won't can Rhapsody in favor of another one of his "cool technologies" the day after you spent $400 for a Rhapsody license?? Sure, the current Macs will continue to function long after Apple ceases to produce them. However, without software vendors continue to develop and update software to keep up with the rest of the world, Macs will quickly become useless. No computers are isolated today. Steve -- Steven Lee Reply address was modified to fight off spammers. Please e-mail me at slee1 "at" erols "dot" com.
From: scott@doubleu.com (Scott Hess) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 1997 05:04:53 GMT Organization: Is a sign of weakness Message-ID: <SCOTT.97Sep9005833@slave.doubleu.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> <5v0h4t$1ir$1@news.lth.se> <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com> In-reply-to: LTaylor7@ix.netcom.com's message of Mon, 08 Sep 1997 11:50:03 -0700 In article <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com>, LTaylor7@ix.netcom.com (Daniel L. Taylor) writes: In article <5v0h4t$1ir$1@news.lth.se>, d89cb@efd.lth.se (Christian Brunschen) wrote: > >NC's have been generally proposed as low-cost diskless stations > >which survive off of a server and lower the per-seat costs for > >corporate computing by lowering the initial cost in addition to > >administration costs (centralized setup, simplified Java-OS). > > Yes. But noone has really suggested that the _CPU_ be crippled as > well. (and adding a local disk should usually be seen as adding a > client-side 'persistent cache', so fits within the concept) The architecture is "crippled" from the outset. No expansion, little RAM, no disk/small disk, all to keep costs down. Things aren't going to move that quickly over Ethernet, even the 100Mb/s variety. Look ahead a couple years, though. Will things move that quickly over 100Mb/s _switched_ Ethernet? How about Gigabit ethernet? How about Gigabit switched ethernet? How about 155Mbit or higher ATM LANs? Of course NCs aren't perfect for _today's_ networking infrastructure - but there's something to be said for being ready with the perfect solution when tomorrow arrives. The way I look at it, we're going to be asking to run videoconferencing and shared visualization tools over this same net. Executables are small compared to a video stream. Later, -- scott hess <scott@doubleu.com> (606) 578-0412 http://www.doubleu.com/ <Favorite unused computer book title: The Demystified Idiots Guide to the Zen of Dummies in a Nutshell in Seven Days, Unleashed>
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 1997 16:38:07 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5v2qsf$dkv$1@attila.apana.org.au> References: <5urae7$f7c$1@attila.apana.org.au> <19970907224459112636@slip22126.rmii.com> <5v04b4$h0a$1@attila.apana.org.au> <alex-0809970817390001@p1-40.van.tvs.net> In article <alex-0809970817390001@p1-40.van.tvs.net>, Alex Curylo <alex@witty.com> wrote: :In article <5v04b4$h0a$1@attila.apana.org.au>, mjm@attila.apana.org.au :(Matt McLeod) wrote: : :> Then perhaps Apple (and other commercial software developers) should learn :> from the freeware experience. : :Um, excuse me here, but the key word is "free". : :The "freeware experience" is a HOBBY, by which I mean "something done for :personal satisfaction which is not necessary to support one's lifestyle". :Observing that lots of hobbyist programmers produce a great amount of code :is about as insightful as observing that lots of hobbyist birdwatchers :produce a great amount of bird sightings. : :If you're trying to develop software in order to feed your family and the :families of the others in your firm, like myself, instead of for your own :personal satisfaction, like the collective Linux community, then it seems :to me there's not too much of value to be learned from this other than the :truism that many hands make light work. : :But if you have any specific suggestions on how to harness hobbyists's :energies like Linus in support of a product that people's families ARE :depending on revenue from to keep food on the table and a roof over their :heads, please share ;) (a) Make the software free. Give away the source. Put Rhapsody under GPL (b) Make the development tools free too. (c) Recruit programmers who are actually interested in the project. Not just people who want a job. (d) Keep management away from the programmers. Set some basic goals, then get the hell out. The company then becomes the maintainer of the software, and still sells it as a shrinkwrap product (nothing in the GPL prevents this). Make your money on shrinkwrap sales to the consumer market, support, and consulting. Netscape went part of the way - giving away betas to anyone who wanted them, which gives their software lots of testing. Take it that extra step and give away the source too, and it wouldn't be so bloated or buggy. Why is OS/2 in such bad shape? Because IBM doesn't care about it, and since it's closed nobody can go in and fix bits themselves. The same is (or could be) true of other software products. OK, it's risky. It's different, which means it would be very difficult to sell to the accountants and managers. But it could work. -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 1997 16:41:07 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5v2r23$dn9$1@attila.apana.org.au> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <gesNOSPAM-0709971424390001@dialup45.sisna.com> <5v04hq$h1p$1@attila.apana.org.au> <alex-0809970824160001@p1-40.van.tvs.net> In article <alex-0809970824160001@p1-40.van.tvs.net>, Alex Curylo <alex@witty.com> wrote: :In article <5v04hq$h1p$1@attila.apana.org.au>, mjm@attila.apana.org.au :(Matt McLeod) wrote: : :(And no, I don't beleive that it's this "NC" :> thing - at least, not the diskless, Java-based, thing). : :However, I wouldn't take any bets that it's not the swapdisk, :xxxStep-based, thing. Indeed, it seems from random c.s.next.* postings :that lots of people DO run their xxxStep networks in pretty much the way :NC networks are envisioned to work -- applications live on the server and :whatever random machine you sit down and log in at, you get access to your :home desktop as if you were there. Which, IMHO, is a good way to do things. If you're paranoid that the sysadmin has the time and inclination to read your letters to Grandma, you can still store them locally. (But if you're that paranoid, you need to get a grip on reality). Matt -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 1997 16:43:53 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5v2r79$ec2$1@attila.apana.org.au> References: <5v04hq$h1p$1@attila.apana.org.au> <B039FDA0-CBE63@206.214.136.134> <5v26mf$fmu@dfw-ixnews11.ix.netcom.com> In article <5v26mf$fmu@dfw-ixnews11.ix.netcom.com>, Tony Invencio <tonyinv@hotmail, .dot, com> wrote: :Benjamin Smith wrote: :> What's going on with the search for a CEO? :> :> Ben S. : :Hey Ben would you want to be CEO around Jobs? : :Think about that before you answer. : :Unless you agree with all of Jobs's moves, I think any person would :think twice before taking that Job. I'd take it. :-) I could use the extra work, and the golden parachute when Jobs gets you kicked out for independent thought is bound to be pretty impressive. -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: Christiaan Hunter <chunter@neumann.une.edu.au> Newsgroups: comp.sys.next.programmer Subject: NSCoder Date: Tue, 09 Sep 1997 16:53:59 +1000 Organization: University of New England, NSW, Australia Message-ID: <3414F287.41C6@neumann.une.edu.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit CC: chunter@neumann.une.edu.au Does anyone have some examples of using NSCoder. I'm having trouble with NSUnarchiver, when decoding my objects. OpenStep manuals only give simple tuts, which I found have a problem. A example with more than two objects being coder and uncodered would be greatly appreciated. That is an object in an object in an Object. Thanks Christiaan, chunter@neumann.une.edu.au...
From: d89cb@efd.lth.se (Christian Brunschen) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 1997 09:04:02 GMT Organization: Lund Institute of Technology, Sweden Message-ID: <5v33e2$3pk$1@news.lth.se> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> <5v0h4t$1ir$1@news.lth.se> <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com> NNTP-Posting-User: d89cb ll have lots of faith in Apple > >Whatever. One look at this newsgroup should be enough to send chills up >Apple's spine.... Yes -- because then they would see that some of their most fervent 'advocates' are ready to abandon the platform based on _rumors_ about what may or may not be happening sometime in the future. Rather chilling an experience indeed. > >> >* Apple appears to be ditching both PowerPC and the desktop, losing four >> >major advantages (Mac OS customer base, Mac OS compatibility - which won't >> >mean anything without a customer base and growing software library, >> >superior price/performance, and a scalable strategy that lets them service >> >a wide range of needs). >> >> Assertations based on rumors, but not on provable fact. > >You know what? I can no longer live by ignoring the clues and just hoping >it will all turn out right. I am not, I am just weighing each rumor, each actual action, weighing them all together, and trying to arrive at a somewhat more balanced picture. >Do you really want me to honestly recommend a >Mac OS system to *anybody* at this point? Yes, of course. They are still the best for many applications. Has that changed because there are rumors that some things might be happening sometime ... ? >Apple has smashed their original >strategy under Amelio, one I put great faith in. Have they? Licensing has changed, but certainly not ended (we have seen three cloners gain MacOS 8 licenses, and now there are these rumors about continued talk between Apple and Motorola); Rhapsody (which was a big part of Amelio's strategy) is coming along better than expected. >And with no clear >strategy to replace it, why on Earth should I stick by? Why should >anyone? Because MacOS is better than ever, and Rhapsody will be a huge leap forward? Because with Rhapsody running on Intel as well as on PPC, the whole 'licensing' controversy becomes somewhat moot? > >Yeah, my theory of what's happening is based on rumor at this point. But >in the absence of a clear strategy that the customer base can evaluate and >base purchasing decisions on, that's all we've got. Given that, you're in >no better position than I am. You *cannot* prove that Apple has a solid >strategy, or even *begin* to evaluate if they will succeed or not. Nor can you _dis_prove it. Also, one doesn't always _need_ a 'solid strategy'. Sometimes, making thigs up as you go along, _works_. > >And that makes Wintel look *very* attractive.... Well, Microsoft _does_ have a 'solid strategy': Market Domination, by more or less any means possible. >> >> I'm not talking about 'Amelio's Plan' , I am talking about Apple's >> _actions_. > >Which have been exceedingly poor and confusing at this point. According to your reading of rumors etc, yes. But not according to mine. >> >> As Apple surely knows -- and therefore they won't 'go for pure NCs'! > >Do they know this? Give me a press statement, or a paper articulating >their strategy. You are the one who made the bold statement that they are going for pure NCs. Unless otherwise proven, I am going to assume that Apple is going to continue to pursue their _current_ strategy. Can _you_ show me a press statement or a paper articulating that Apple is going to switch to 'pure NCs' ? > >Don't have one? Hmmm.... Well, I admire you for not giving in to rumors as >I have (I admit this). But you know what? At least the rumors keep me >interested and create a period of pause before I abandon the platform. If >I were to make a cold, purely logical, based 100% on fact purchasing >decision at this moment, I would be forced to default to Wintel in the >absence of any stated plan on Apple's part to survive. I follow the rumors closely. I just don't give in to them; I take them into account, but try to see beyond them. Rumors are just that: rumors. Apple _has_ a 'stated plan to survive': Rhapsody. For me, who knows OPENSTEP, that is more than sufficient incentive to stay with Apple. Just for the technical merits of the system. (Much like BeOS advocates stay with Be, who have approximately as much of a strategy as Apple has. And I am certaonly not counting Be out of the race.) > >In other words, Apple isn't even on the playing field right now. > >Is that preferable? Oh, but they are -- it's just not necessarily the playing field you are used to. > >> >Power Computing is being bought by Apple and there is no point in denying >> >them OS 8. >> > >> >Power Tools may just be shipping the box...we don't know yet. Power Tools have announced they will ship with MacOS 8 preinstalled, and included on CD. >> > >> >UMAX just has OS 8, no CHRP. At the moment, yes -- but according to rumor (which you seem to listen so closely to), UMAX is talking with Apple about future licensing. Which might well include CHRP. >> > >> >Motorola is screwed. Again, according to the same rumor, Motorola are once more talking with Apple about licensing. Which might include CHRP. >> > >> >There are no other cloners. >> >> Counterexample: PowerTools. Who are, as we speak, _shipping_ MacOS >> clones with MacOS 8, with Apple's blessing. (I mentioned them above, >> too.) > >Oh well, forgive me, *that's* an open market... :-/ Well, 3 cloners is a _lot_ more than _no_ cloners. Also, if the rumors are true, we are just seeing the beginning of the 'new licensing' from Apple, which will replace the old licensing (which Apple won't extend). So there may well be more cloners, talking with Apple. >> >> You basically claim that Apple's recent actions have pissed off _all_ >> of Apple's customers, so that they won't buy any new Apple >> products. > >Not all. Most. Fortunately, you are still just as wrong :) >> >> That largely remains to be seen - and, of course, opinions >> differ. Many people actually think that Apple's recent decisions are >> _good_ ones. > >Who? And no, Larry Ellison and Steve Jobs do not count.... From: nospam1933@aol.com (Nospam1933) Newsgroups: rec.backcountry Subject: How do spammers get our email address? Date: 9 Sep 1997 09:03:45 GMT Lines: 9 Message-ID: <19970909090300.FAA12917@ladder01.news.aol.com> NNTP-Posting-Host: ladder01.news.aol.com X-Admin: news@aol.com Organization: AOL http://www.aol.com SnewsLanguage: English Path: genius.dat.hk-r.se!news.lth.se!eru.mt.luth.se!feed1.news.erols.com!howland.erols.net!news-peer.gsl.net!news.gsl.net!gip.net!portc01.blue.aol.com!newstf02.news.aol.com!audrey01.news.aol.com!not-for-mail Does anyone know how those god-condemned spammers do their thievery? Do they use programs to automatically steal email addresses from newsgroup posts? Or do they manually search for valid email addresses? From my past experience, I seem to get no junk mail sent to spam-unfriendly screen names such as "nospam@xyz.com" even when I set my email address to accept all email. This could mean that the spammers use programs that ignore spam-unfriendly screen names. Or it could also mean that they manually copy email addresses and ignore those that sound spam unfriendly. $TW-XK71 M'4`+_DM$2K+5FXE][6Q!%P.,<('&VJBUB@72>OG_/`>FG5/\@P&-\[O(#:&W M^3@9*?EJR-HY`M1=5+N_D8"9P1&S3B"\/"L640F"P?\\C>FS&HZV(NT)`7_2 M1`]'IB=P<QONH@%,+UY>L,7F;8.8T:)EVUH`-4;7]Q$`8=*E^I9BV@)]<=)! MJ.IZN;JD%4PU\D3@BG*`A$ZZ1ZF+"Y\7V<3"KP[#1+()QX87CYR?LQT3>P!6 M-_-W"R`S4$CY+2RA\BJDB%J''QC4;X32T\].ZCV=^JYD&JNPMD*3^K7WOM[J M?.<[U=!R/7ORG\$`%,F0EEDMRSW/-;WZ3\[4U<)U.6$K!F)>.``USQ<=`0&Q MDKM$]O$(>)7:-7SHAJ*FVUO0;<.--RU#O``+?*^N?M>_:;GG>/=]CX3Q#_KZ MB+[XW=`/_3RCFG(`]#9\K=E:BZ#&!`1N)V``.0V@6!]0D+\\P7\"%Y?^^%0S M:"9^^_$&(XQ]P%KSL=4AM8`:]V].7H$58@$`G<,[OSH/_ZG[A.(GB-`1Y/EY M9,;S-9<80FQMIARM-1FR0'@>X+98N^(U1SGRT`T@&(7A9H47],=C+N%`D2</ M8XZH"E#6H,6O>7<+X#M02/@M+.'R*;B<Z7[:,Z#59B6RTXKJUY3V?)?:>PZ; M.E)ME/%7GZETIZWK5*ZETXG6S2FB"7WG^Q>\33TUFJHT*MW\=.A+U3!,/_59 MVBG"U!"ER<L!WW%.$0"$9HY2%,KA%.G_&,UM^*@Q:[$U7T"MH@N(.XE"$87- MR#=)35^QY;-QQE1MJGRMM=;1`(^$]=#LS1H1K^1FFD*UUFE"Y.&:P09@\P+4 M%+'WB->X&NMZ9H`W]'\(R\@7:_R/B?!J4&L$$'-(KI;]@2?E&5*/EGA?X>4X M<7LTB2.&RB>!".5G3#/<*YQ-H<M-%MQ;U.VVM%;P#=$-%P$!GV>$.2,WF_A* M:>S-,7GU0B@!QP_X39H`;;NR=PN.LU!(^"TLX?(IOI*=4CDL:<XYT0("-&3" MWEU=.RN?+E.F4@VZ9\#22DOX$])4?JTRZ0_9/FFFA'B5GTI+MB0-Z#?2E`>E M)URARWR^^JGO6BJ?^7S)RP:K=)%2#T9[:XGVVJTJP:*F$,KTW0`F19Y4V;#V M%V8R4AH&HDTR0Y8!+32Z"$9=_]5\G)QSZ4)<29:M7EIK$>@9C/L:+4-4<YS6 M;.V`&M<[*3-'*<_Y#%Y.!VC$[3;[^3C;A^KE7J2M9W.V!K#-U@*HJ5P"/"IS MY>&AFB%LJ5B8;0F9WU5)VYSCA6)4V790K);7I9%::X"`U@F$4T&2-OQ&,A91 ML\LLL/'$8/.K5ESLTM06,LEB/CB%CM":OVMW"S_:4$CX+2SA\BH@Q$:'][O2 M3DO3+`W*-T-IIDREPZDPAD*S^M3GO]S]<N<SWM3RPU72[\'WP?LQ!.;.YT$R MVBK7CL&%MDTYM8/3"UTP9TS3V]9K4`10X-DHXG?>8,7B_$O+-0V";D*G"#?, M"5.OX0@EO!N$D89`[?$8I#L'5.\0-;_\SU0T&M=79%LCK`M9]G3.YBP,LIJS M-=B#`(;0/XNN[-@#Q'FFVY#B4N)$)):,VC"VU[-M52:MP*!J!-0`"NN3PP7" M9<$*YP4$G9LJ:*$VF3V5YKC:VJ&:"K9:?5G;2XK62*BU(`A3`'*((;M>,\!L M'35/).^^7K]/1XU42+Z-E3:Y!Y)@8T+29<KFCJV10&O.4G<+Q#M02/@M+.'R M*CF:Y7[T^QV<2ROT"0]$0ZOH5IN*88G-O?)-K[12G_U\V-.K;[*Z5M(IPH__ M%+U_&#CA$?6'_6GK(P]T*/E4/KOF:XT\6E.UIFNME&RT`OMO&*!Q'T1@_?2$ M[W=46)'QJID:0+96From: d89cb@efd.lth.se (Christian Brunschen) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,
Date: Tue, 09 Sep 1997 05:41:18 -0600 From: David Stes <stes@can.nl> Subject: Re: Exception processing in Objective C? Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Message-ID: <873800789.32459@dejanews.com> Organization: Computer Algebra Netherlands References: <340C6C07.5727@win.bright.net> <5v1f07$loh@shelob.afs.com> Rich Schroedel <richs@win.bright.net> writes > I'm in the process of rewriting some C++ code so that it can eventually > be converted to Objective C. I'm stumped at how to handle exception > processing. How do Objective C programmers handle the sort of situations > that C++ programmers handle with exception processing? Any help out > there? I think this can best be done by using Objective-C Blocks. For example, to handle "out of memory" in +new, (if you would not use -gc), [{ myObject = [Object new]; /* get an instance */ } ifException: { id exception | /* Block with one argument */ [uselessObjects free]; /* free some memory */ [exception resume]; /* "resume at location where failed */ }]; This (specific) exception handler actually doesn't long-jump out of +new, because of the |resume| message. If malloc() (inside +new) returns NULL, then the handler will be evaluated. This particular handler frees some memory, then +new resumes, and will attempt to malloc() again (see below) until it succeeds. Of course, the handler could, instead of recovering, just fall through (ie. long jump towards -ifException:) or reraise the exception, which would evaluate the default handler, a Block that sends an -error: message (aborts the process) saying "out of memory in +new" : + new { id newObject; static id exBlock; /* exBlock needs to be instantiated before we're running out memory */ if (!exBlock) exBlock = [{ [self error:"out of memory in +new"] } self]; while ((newObject = (*_alloc)(self,0)) == nil) [exBlock raise]; return newObject; } You can use Blocks for other things as well, like: id filtered; id aCltn = [Collection new]; ... filtered = [aCltn collect:{id e | [set find:e]}]; which builds a Collection where it's rejecting the elements that it does not find in "set", but anyhow, exception handling is probably the most serious application. David. -------------------==== Posted via Deja News ====----------------------- http://www.dejanews.com/ Search, Read, Post to Usenet
From: Arnaud Debayeux <debayeux@easynet.fr.NO_SPAM> Newsgroups: comp.sys.next.programmer Subject: Custom message dispatching ? Date: Tue, 09 Sep 1997 14:10:27 +0200 Organization: Systems Group, Easynet Ltd. Message-ID: <34153CB3.FB6BA967@easynet.fr.NO_SPAM> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, Is it possible in NextStep/OpenStep to implement a custom message dispatching by modifying ONLY the basic implementation ? I want to avoid to add any code to the objects. Thanks
From: Tony Invencio <"tonyinv"@hotmail .dot com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 08:59:18 -0400 Organization: Have None Message-ID: <5v3h76$ic@dfw-ixnews7.ix.netcom.com> References: <5v04hq$h1p$1@attila.apana.org.au> <B039FDA0-CBE63@206.214.136.134> <5v26mf$fmu@dfw-ixnews11.ix.netcom.com> <5v2r79$ec2$1@attila.apana.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Matt McLeod wrote: > > In article <5v26mf$fmu@dfw-ixnews11.ix.netcom.com>, > Tony Invencio <tonyinv@hotmail, .dot, com> wrote: > :Benjamin Smith wrote: > :> What's going on with the search for a CEO? > :> > :> Ben S. > : > :Hey Ben would you want to be CEO around Jobs? > : > :Think about that before you answer. > : > :Unless you agree with all of Jobs's moves, I think any person would > :think twice before taking that Job. > > I'd take it. :-) > > I could use the extra work, and the golden parachute when Jobs gets > you > kicked out for independent thought is bound to be pretty impressive. One of his first actions was gettin rid of those "Golden Parachutes". -- Tony Invencio Strange Solutions Software "The fault, dear Brutus, is not in our stars, But in ourselves, that we are underlings." Cassius
Date: Tue, 09 Sep 1997 11:31:09 -0400 From: joe.ragosta@dol.net (Joe Ragosta) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Message-ID: <joe.ragosta-ya02408000R0909971131090001@news.dol.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <5up1dc$ndg$1@news.lth.se> <341074F1.8F26E657@spamtoNull.com> <5urae7$f7c$1@attila.apana.org.au> <gesNOSPAM-0709971424390001@dialup45.sisna.com> <nagleEG77Bp.Kor@netcom.com> Organization: Graver Chemical Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit In article <nagleEG77Bp.Kor@netcom.com>, nagle@netcom.com (John Nagle) wrote: > gesNOSPAM@sisna.com (ges) writes: > >All this rationalizing doesn't reassure me. I think Apple is committing > >suicide, or rather Jobs is murdering it. > > Apple should go out of business while they still have a positive > net worth. They should just say "we lost - it's over", instead of > this death march to bankruptcy. The stockholders would still get something. > As it is, they may lose everything. Nonsense. As a stockholder, I don't want them to go out of business. If they wanted to go out of business, stockholders would get a lot more if they sell the company than if they close the doors and sell off the remaining inventory. -- Regards, Joe Ragosta See the Complete Macintosh Advocacy Site http://www.dol.net/~Ragosta/complmac.htm
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 09:11:58 -0700 Organization: Netcom Message-ID: <LTaylor7-0909970911590001@vic-ca1-11.ix.netcom.com> References: <5v04hq$h1p$1@attila.apana.org.au> <B039FDA0-CBE63@206.214.136.134> In article <B039FDA0-CBE63@206.214.136.134>, "Benjamin Smith" <benjs@ix.netcom.com> wrote: > The thing of it is is that the current people on Apple's board are proven. > Many of them are from the industry. So, they have a plan that they think is > sound. I guess many people believe that Jobs is calling all shots. I don't. > > And maybe they have another vision and this just is not the time to > announce it. And there may be some very sound reasons behind it. And in the mean time, they are losing sales. > I just wish that people would at least entertain this notion and not the > familiar one that Jobs is trying to destroy Apple or he is destroying it > through a series of wrong-headed moves. I think it is too early to tell. I can't "entertain" clients in this manner. People have purchase decisions coming up. I cannot recommend Mac OS based on "Well, *maybe* Steve and the BOD have a great plan to save Apple...even though thus far they've smashed what was a great plan and lied to their customer base about cloning...put your faith and $$$ in them any way...". -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 09:15:08 -0700 Organization: Netcom Message-ID: <LTaylor7-0909970915090001@vic-ca1-11.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> <5v0h4t$1ir$1@news.lth.se> <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com> <SCOTT.97Sep9005833@slave.doubleu.com> In article <SCOTT.97Sep9005833@slave.doubleu.com>, scott@doubleu.com (Scott Hess) wrote: > Look ahead a couple years, though. Will things move that quickly over > 100Mb/s _switched_ Ethernet? How about Gigabit ethernet? How about > Gigabit switched ethernet? How about 155Mbit or higher ATM LANs? Of > course NCs aren't perfect for _today's_ networking infrastructure - > but there's something to be said for being ready with the perfect > solution when tomorrow arrives. Perfect solutions require perfect business decisions and perfect marketing and perfect PR. Neither of which Apple is demonstrating right now. -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 9 Sep 1997 03:38:03 GMT Organization: Genoa Software Systems Message-ID: <5v2gar$8e@saturn.genoa.com> References: <5uklcd$dmp$1@concorde.ctp.com> <5ul4sv$llm$1@news.apple.com> <5um35u$sg5$1@concorde.ctp.com> <x7d8mpcd8i.fsf@tom.ics.ele.tue.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: tiggr@ics.ele.tue.nl In <x7d8mpcd8i.fsf@tom.ics.ele.tue.nl> Pieter Schoenmakers wrote: > Messaging nil is a bad idea. --Tiggr One man's opinion. My experience is that allowing messages to nil, removes alot of clutter from code, making it more clear and readable. [foo bar]; is exactly equivalent to the more verbose if (foo) [foo bar]; You just have to realize that the (relatively rare) case where the return point is not void, int or a pointer requires extra thought since the return value is undefined in that case. Let's not start a religous war, I just wanted to state that there is another point of view on this issue. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 09:25:52 -0700 Organization: Netcom Message-ID: <LTaylor7-0909970925550001@vic-ca1-11.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> <5v0h4t$1ir$1@news.lth.se> <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com> <5v33e2$3pk$1@news.lth.se> In article <5v33e2$3pk$1@news.lth.se>, d89cb@efd.lth.se (Christian Brunschen) wrote: > ll have lots of faith in Apple > > > >Whatever. One look at this newsgroup should be enough to send chills up > >Apple's spine.... > > Yes -- because then they would see that some of their most fervent > 'advocates' are ready to abandon the platform based on _rumors_ about > what may or may not be happening sometime in the future. Rather > chilling an experience indeed. The end of open cloning is *not* a rumor. As for the rest, I cannot blame any one here. Apple has lied to its customer base about cloning, and now leaves an information void. If Apple's PR can't handle filling that void, then rumor is all the Mac base has. And that's a good indication of just how dependable Apple will be in other matters. > >Do you really want me to honestly recommend a > >Mac OS system to *anybody* at this point? > > Yes, of course. They are still the best for many applications. Has > that changed because there are rumors that some things might be > happening sometime ... ? No, that has changed because the Mac market is now closed and Wintel is rapidly taking back price/performance points. A closed market casts doubt upon the future survival of Mac OS, while Apple smashing yet another strategy yet not having a replacement does nothing to ease that doubt. I see lame-brained business moves and PR all over again, and that's what changes my recommendation. > >Apple has smashed their original > >strategy under Amelio, one I put great faith in. > > Have they? Licensing has changed, but certainly not ended It has *ended* by my definition. I do not consider letting a few cloners limp along in a restricted way to be any form of open market. > (we have > seen three cloners gain MacOS 8 licenses, and now there are these > rumors about continued talk between Apple and Motorola); Rhapsody > (which was a big part of Amelio's strategy) is coming along better > than expected. Really? Was the DR not to be out in the summer? Is it just me, or did it slip? > >And with no clear > >strategy to replace it, why on Earth should I stick by? Why should > >anyone? > > Because MacOS is better than ever, and Rhapsody will be a huge leap > forward? Because with Rhapsody running on Intel as well as on PPC, the > whole 'licensing' controversy becomes somewhat moot? If Apple can so flippantly abandon CHRP and clones, why in the world should I trust them on Rhapsody or Mac OS? > >Yeah, my theory of what's happening is based on rumor at this point. But > >in the absence of a clear strategy that the customer base can evaluate and > >base purchasing decisions on, that's all we've got. Given that, you're in > >no better position than I am. You *cannot* prove that Apple has a solid > >strategy, or even *begin* to evaluate if they will succeed or not. > > Nor can you _dis_prove it. Also, one doesn't always _need_ a 'solid > strategy'. Sometimes, making thigs up as you go along, _works_. Tell that to my friends and clients.... > >> I'm not talking about 'Amelio's Plan' , I am talking about Apple's > >> _actions_. > > > >Which have been exceedingly poor and confusing at this point. > > According to your reading of rumors etc, yes. But not according to > mine. Killing CHRP, ending cloning "as we know it" (an Apple Recon popular phrase), and handling it all in a poor manner is *not* rumor. > >> As Apple surely knows -- and therefore they won't 'go for pure NCs'! > > > >Do they know this? Give me a press statement, or a paper articulating > >their strategy. > > You are the one who made the bold statement that they are going for > pure NCs. Unless otherwise proven, I am going to assume that Apple is > going to continue to pursue their _current_ strategy. Which is what? > >Don't have one? Hmmm.... Well, I admire you for not giving in to rumors as > >I have (I admit this). But you know what? At least the rumors keep me > >interested and create a period of pause before I abandon the platform. If > >I were to make a cold, purely logical, based 100% on fact purchasing > >decision at this moment, I would be forced to default to Wintel in the > >absence of any stated plan on Apple's part to survive. > > I follow the rumors closely. I just don't give in to them; I take them > into account, but try to see beyond them. Did you read what I said? > >> You basically claim that Apple's recent actions have pissed off _all_ > >> of Apple's customers, so that they won't buy any new Apple > >> products. > > > >Not all. Most. > > Fortunately, you are still just as wrong :) In my experience, most. I'm already losing 3 people to Wintel as we speak. And no, I'm not as negative about things with them. I simply tell them what's happening (they have purchasing decisions coming up), and they are pretty much going NT. -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: LTaylor7@ix.netcom.com (Daniel L. Taylor) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 09:29:17 -0700 Organization: Netcom Message-ID: <LTaylor7-0909970929170001@vic-ca1-11.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <LTaylor7-070997 <5v1etj$84r$1@news.fas.harvard.edu> In article <5v1etj$84r$1@news.fas.harvard.edu>, ibhan@digitas.harvard.edu (Ishir Bhan) wrote: > LTaylor7@ix.netcom.com (Daniel L. Taylor) writes: > > >My point isn't that NC's are bad, it's that ditching the Mac market and > >going for pure NC's will kill Apple. > > Why do you always equate these two? Why do you think that going for NC's > will mean ditching the Macintosh market (unless you think that eventually > switching people over to Rhapsody will be "ditching the Mac market"). > Apple will almost sell both types of machines. You are correct, and I've already apologized for this. This was my theory as to the going-on's of Steve's brain. I don't tell clients or friends this when discussing Apple's situation, I guess I shouldn't help spread rumors here either. But Steve's Fortune quote has me very doubtful and worried about his plans. He stated that if he were in control he would milk the Mac for all it's worth and move on to the next great thing. It's no secret that he and Ellison believe NC's are the next great thing. No, I shouldn't argue from my own personal hunch based on one Steve Jobs quote. But often you don't have to look far to figure out what a person is planning. Just read their lips.... -- Daniel L. Taylor LTaylor7@ix.netcom.com ================================================================= The Heritage Foundation /// Town Hall http://www.heritage.org/heritage/ /// http://www.townhall.com/ =================================================================
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 9 Sep 1997 16:45:17 GMT Organization: Anderson Financial Systems Inc. Message-ID: <5v3uet$qkf@shelob.afs.com> References: <5v2gar$8e@saturn.genoa.com> Alex Blakemore <alex@genoa.com> writes > In <x7d8mpcd8i.fsf@tom.ics.ele.tue.nl> Pieter Schoenmakers wrote: > > Messaging nil is a bad idea. --Tiggr > > My experience is that allowing messages to nil, removes alot of clutter > from code, making it more clear and readable. > > [foo bar]; > is exactly equivalent to the more verbose > if (foo) [foo bar]; I happen to agree with Alex's point of view. The one gotcha is that _this_ construction: float value = [aField floatValue]; returns different values on different architectures when aField == nil. On m68k, value = 0 after this message. In i386, it can be NaN. This is clearly an undesirable result. So you do have to be careful about sending messages to nil objects that return certain types of values. -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
From: "Benjamin Smith" <benjs@ix.netcom.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 97 13:31:55 -0500 Organization: Netcom Message-ID: <B03B004F-499EDE@199.35.216.140> References: <5v26mf$fmu@dfw-ixnews11.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit nntp://nntp.ix.netcom.com/comp.sys.next.advocacy, nntp://nntp.ix.netcom.com/comp.sys.mac.programmer.misc, nntp://nntp.ix.netcom.com/comp.sys.next.programmer, nntp://nntp.ix.netcom.com/comp.sys.mac.system, nntp://nntp.ix.netcom.com/comp.sys.mac.misc On Mon, Sep 8, 1997 7:53 PM, Tony Invencio <mailto:"tonyinv"@hotmail .dot com> wrote: >Hey Ben would you want to be CEO around Jobs? > >Think about that before you answer. > >Unless you agree with all of Jobs's moves, I think any person would >think twice before taking that Job. > No. Whoever is thinking of this position should have a talk with Ellen Hancock. Ben S.
From: mtrombin@ix.netcom.com (Mark Trombino) Newsgroups: comp.sys.next.programmer Subject: Re: NSCoder Date: 9 Sep 1997 17:56:30 GMT Organization: Egghead Billy, Inc. Message-ID: <5v42ke$4hg@dfw-ixnews11.ix.netcom.com> References: <3414F287.41C6@neumann.une.edu.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: chunter@neumann.une.edu.au In <3414F287.41C6@neumann.une.edu.au> Christiaan Hunter wrote: > Does anyone have some examples of using NSCoder. > > I'm having trouble with NSUnarchiver, when decoding my objects. > > OpenStep manuals only give simple tuts, which I found have a problem. > > A example with more than two objects being coder and uncodered would > be greatly appreciated. That is an object in an object in an Object. Don't know if this will help, but here's a couple of methods that I've used... Let me know if this helps. -- |-------------------------------------------------------------------------| | Mark Trombino | J A M S o f t | | mtrombin@ix.netcom.com | Audio DSP Tools for Openstep & Rhapsody | |-------------------------------------------------------------------------| - (void)saveFile:(NSString *)filename { NSArchiver *newArchive; int index; newArchive = [[NSArchiver alloc] initForWritingWithMutableData:[NSMutableData data]]; if(newArchive == NULL) { NSRunAlertPanel(@"File i/o Error", @"Unable to open typed stream for file: %@", @"Okay", nil, nil, saveFile); return; } [newArchive encodeRootObject:fromLocations]; [newArchive encodeRootObject:toLocations]; [newArchive encodeRootObject:optionLocations]; [newArchive encodeValueOfObjCType:"i" at:&secondsInClock]; index = [fromPopup indexOfSelectedItem]; [newArchive encodeValueOfObjCType:"i" at:&index]; index = [toPopup indexOfSelectedItem]; [newArchive encodeValueOfObjCType:"i" at:&index]; index = [optionPopup indexOfSelectedItem]; [newArchive encodeValueOfObjCType:"i" at:&index]; // contents of typedStream must explicitly be written to a file [[newArchive archiverData] writeToFile:saveFile atomically:YES]; [window setDocumentEdited:NO]; } - (void)openFile:(NSString *)filename { NSUnarchiver *newUnarchiver; int index; newUnarchiver = [[NSUnarchiver alloc] initForReadingWithData:[[NSData alloc] initWithContentsOfFile:filename]]; [saveFile release]; saveFile = [filename retain]; // Release the arrays [fromLocations release]; [toLocations release]; [optionLocations release]; // Decode the arrays fromLocations = [[newUnarchiver decodeObject] retain]; toLocations = [[newUnarchiver decodeObject] retain]; optionLocations = [[newUnarchiver decodeObject] retain]; [fromLocations makeObjectsPerform:@selector(setController:) withObject:self]; [toLocations makeObjectsPerform:@selector(setController:) withObject:self]; [optionLocations makeObjectsPerform:@selector(setController:) withObject:self]; // set the clock [newUnarchiver decodeValueOfObjCType:"i" at:&index]; [self setTime:index]; // select the items that were current in each popup when file was saved [self syncPopups]; [newUnarchiver decodeValueOfObjCType:"i" at:&index]; [fromPopup selectItemAtIndex:index]; [newUnarchiver decodeValueOfObjCType:"i" at:&index]; [toPopup selectItemAtIndex:index]; [newUnarchiver decodeValueOfObjCType:"i" at:&index]; [optionPopup selectItemAtIndex:index]; [window setDocumentEdited:NO]; }
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: Custom message dispatching ? Date: 9 Sep 1997 19:01:47 GMT Organization: Apple Computer, Inc. Message-ID: <5v46er$rru$1@news.apple.com> References: <34153CB3.FB6BA967@easynet.fr.NO_SPAM> Arnaud Debayeux <debayeux@easynet.fr.NO_SPAM> writes > Hi, > > Is it possible in NextStep/OpenStep to implement a custom message > dispatching by modifying ONLY the basic implementation ? I want to > avoid to add any code to the objects. > > Thanks I'm not sure I understand your question, but if you want to change the way that messages are dispatched, you can write your own versions of the Objective-C runtime functions. This is tricky if you only want to change *some* of the functionality. The question would be, what exactly are you trying to do? You might be able to use a higher-level mechanism to achieve the same goals. -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: cdstrand@onramp.net (Dave) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 14:17:09 -0500 Organization: Could be better Message-ID: <cdstrand-0909971417100001@cpsidfw.flash.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <don_arb-0509971214210001@sea-ts1-p59.wolfenet.com> In article <don_arb-0509971214210001@sea-ts1-p59.wolfenet.com>, don_arb@wolfenet.com (Don Arbow) wrote: (( cuts )) > Most of the hype about NC's deals with corporations and large intranets > where employees are using $2500 PC workstations to play solitaire. The > main push behind using NC's is that these companies can get rid of those > pricy machines, if the users are only going to run a word processor and a > database client. Centralized storage for programs and data makes sense > for maintenance and administration costs. What's your take on Why, if virtually every large company has this option now, ie, they can replace their desktop computers for IBM 3274 dumb terminals which are much cheaper, and, yet, they keep on doing the opposite? I mean all you are doing is describing an SNA/SDLC network here. -- Dave cdstrand@onramp.net
From: alex@witty.com (Alex Curylo) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 12:56:42 -0700 Organization: Bavarian Illuminati Message-ID: <alex-0909971256420001@p2-03.van.tvs.net> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> <5v0h4t$1ir$1@news.lth.se> <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com> In article <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com>, LTaylor7@ix.netcom.com (Daniel L. Taylor) wrote: > > Is there any proof that Apple 'has dropped the PowerPC' ? > > This is not Apple's choice, but IBM and Motorola's. Based on how they've > been screwed by Apple, it wouldn't be a surprise. Errrr ... if IBM/Moto do drop PPC desktop CPU development, as does appear likely, I very sincerely doubt the latest jacking around from Apple would be anything except the straw that broke the camel's back. Why don't you whine about IBM (OS/2), Novell (NetWare), Sun (Solaris), and Microsoft (NT) screwing the PPC? Yes, the PPC desktop family has turned out to be bordering on disaster as an investment for IBM/Moto (at least I get that impression, does anyone have real numbers handy on investment and returns??), but it's the fault of the four aforementioned, not Apple. ---- Alex Curylo -- alex@witty.com
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <5v4gpo$3ij@bgtnsc03.worldnet.att.net> Control: cancel <5v4gpo$3ij@bgtnsc03.worldnet.att.net> Date: 09 Sep 1997 21:58:17 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5v4gpo$3ij@bgtnsc03.worldnet.att.net> Sender: JOHN EDMONDS<ANYWHERE@USA.COM> Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: fcrary@rintintin.Colorado.EDU (Frank Crary) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 1997 01:36:32 GMT Organization: University of Colorado, Boulder Message-ID: <5v2970$719@peabody.colorado.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <acurylo-0509971010020001@p2-29.van.tvs.net> <5ur2uv$r3s@crcnis3.unl.edu> <5uujej$2to$1@nntp.ucs.ubc.ca> Return-Path: news@peabody.colorado.edu Delivery-Date: Mon, 08 Sep 1997 18:57:18 -0700 Return-Path: news@peabody.colorado.edu for <gherbert@crl.com>; Mon, 8 Sep 1997 18:31:36 -0700 (PDT) env-from (news@peabody.colorado.edu) id TAA07211; Mon, 8 Sep 1997 19:36:33 -0600 (MDT) To: gherbert In article <5uujej$2to$1@nntp.ucs.ubc.ca>, Tim Shoppa <shoppa@alph02.triumf.ca> wrote: >>: Not to mention the Mars Rover. (PPC -- 100% of the >>: Martian computing installed base! There's an ad for you huh?) >>Not to rain on your extraterrestial parade or anything, but >>doesn't the Mars Rover use a version of the Intel 8085? >Yes, a 80C85 is in the Rover. The lander has a IBM RAD6000, which >is vaguely related to the PowerPC (but it is *not* a PowerPC). Two comments. First, I though it was a RAD4000, which would be a radiation hardened version of the R4000 chip. Second, last time I checked, the PowerPC was a computer (or series of computers), not a chip. So the connection to PowerPCs is simply that they use the a similar processor. (Nor are PowerPC the only computers that use this family of chips: SGI uses them extensively; the Indigo II on my desk at work uses a R4400 processor.) I think it's important to avoid confusing chips with computers. Frank Crary CU Boulder
From: davewang%wam.umd.edu.delete.after.edu@Glue.umd.edu (David T. Wang) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: CPUs in Space (Was Re: No more PowerPCs... Followup-To: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Date: 9 Sep 1997 04:26:35 GMT Organization: University of Maryland, College Park Message-ID: <5v2j5r$tqp$1@hecate.umd.edu> References: <5uujej$2to$1@nntp.ucs.ubc.ca> <B039C5D9-1E794@206.165.44.76> Delivery-Date: Mon, 08 Sep 1997 21:48:54 -0700 Return-Path: news@hecate.umd.edu for <gherbert@crl.com>; Mon, 8 Sep 1997 21:22:13 -0700 (PDT) env-from (news@hecate.umd.edu) by hecate.umd.edu (8.8.7/8.8.7) id AAA28176; Tue, 9 Sep 1997 00:26:38 -0400 (EDT) To: sci-space-tech@moderators.uu.net Lawson English (english@primenet.com) wrote: : Tim Shoppa <shoppa@alph02.triumf.ca> : > Yes, a 80C85 is in the Rover. The lander has a IBM RAD6000, which : > is vaguely related to the PowerPC (but it is *not* a PowerPC). : > : > Do we really want these sandbox-type "my CPU is better than your CPU" : > arguments spilling over into sci.space.tech? : > : Just curious. The more modern CPUs with their denser transistor count, seem : like they would be more vulnerable to things like errors due to cosmic : rays. What kind of trade-off between vulnerability (if there is one), : CPU-power, power consumption, size, etc., leads to the selection of a given : CPU for such a mission? You are ofcourse correct. From what I understand, if you use one of the older process technologies, like 2.0um, it has thick enough oxides so that you won't easily get your oxides annihilated by cosmic ray. However, there's not much you can do about single event upsets. (Cosmic ray goes through, and charges/discharges a single bit, and flips it.) So they don't use any sort of SRAM based FPGA's kind of thing out in space. The funny thing is that while we're enjoying 0.25um down here on earth, most of the CPU's flying out into space are the ancient 8 bit or 16 bit stuff, built on Radiation hardened process that's like 2.0um or smth big. AFAIK, most of the CPU's actually flying in outer space are, by our standards, really old stuff. We're sending 6 486's, but only to the edge of space on a huge balloon for a few weeks, not Mars for a few months/years. : ------------------------------------------------------------------------- : Build a better mousetrap, yes, but then market it like crazy, or instead of : catching mice you'll only collect dust. -David Yeargin : -------------------------------------------------------------------------
From: bmeadows@acl1.physics.gatech.edu (Brian K. Meadows) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: CPUs in Space (Was Re: No more PowerPCs... Date: Tue, 09 Sep 1997 14:10:26 GMT Organization: Georgia Institute of Technology Message-ID: <341556f5.560572@news.gatech.edu> References: <5uujej$2to$1@nntp.ucs.ubc.ca> <B039C5D9-1E794@206.165.44.76> Delivery-Date: Tue, 09 Sep 1997 07:10:48 -0700 Return-Path: news@smash.gatech.edu for <gherbert@crl.com>; Tue, 9 Sep 1997 07:06:54 -0700 (PDT) env-from (news@smash.gatech.edu) id QQdggm13499; Tue, 9 Sep 1997 10:11:35 -0400 (EDT) On 8 Sep 1997 14:48:00 -0700, "Lawson English" <english@primenet.com> wrote: > >Tim Shoppa <shoppa@alph02.triumf.ca> > >> Yes, a 80C85 is in the Rover. The lander has a IBM RAD6000, which >> is vaguely related to the PowerPC (but it is *not* a PowerPC). >> >> Do we really want these sandbox-type "my CPU is better than your CPU" >> arguments spilling over into sci.space.tech? >> > >Just curious. The more modern CPUs with their denser transistor count, seem >like they would be more vulnerable to things like errors due to cosmic >rays. What kind of trade-off between vulnerability (if there is one), >CPU-power, power consumption, size, etc., leads to the selection of a given >CPU for such a mission? There are several know "process receipts" for CMOS, BJT, and Bi-CMOS transistors that include hardening against radiation. Rad-hardened chips are made in quantity for the military. Engineers for Pathfinder could always use shielding, but I doubt the designers wanted to pay the weight penalty. Brian
From: beatty@netcom.com (Derek Lee Beatty) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 7 Sep 1997 22:24:53 GMT Organization: none Message-ID: <5uv9jl$5qh@beatty.slip.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <acurylo-0509971010020001@p2-29.van.tvs.net> <5ur2uv$r3s@crcnis3.unl.edu> Delivery-Date: Mon, 08 Sep 1997 22:39:23 -0700 Return-Path: news@news9.noc.netcom.net for <gherbert@crl.com>; Mon, 8 Sep 1997 22:31:38 -0700 (PDT) env-from (news@news9.noc.netcom.net) id QQdgfe00225; Tue, 9 Sep 1997 01:36:21 -0400 (EDT) id AAA11005; Tue, 9 Sep 1997 00:36:00 -0500 (CDT) To: sci-space-tech@uunet.uu.net 00093182@bigred.unl.edu (Josh Hesse) wrote: >Not to rain on your extraterrestial parade or anything, but >doesn't the Mars Rover use a version of the Intel 8085? > >I could almost swear I saw it in _Aviation Week_ or someplace. The Mars lander has an IBM RS/6000 on board, and communicates with the rover using a Motorola radio modem. The rover might have an 8085 on board. -- Derek Lee Beatty _ Death beatty@netcom.com _| ~-, Taxes Austin, Texas \, * } C++ \_(
From: scott@doubleu.com (Scott Hess) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 1997 22:05:06 GMT Organization: Is a sign of weakness Message-ID: <SCOTT.97Sep9174449@slave.doubleu.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> <5v0h4t$1ir$1@news.lth.se> <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com> <SCOTT.97Sep9005833@slave.doubleu.com> <LTaylor7-0909970915090001@vic-ca1-11.ix.netcom.com> In-reply-to: LTaylor7@ix.netcom.com's message of Tue, 09 Sep 1997 09:15:08 -0700 In article <LTaylor7-0909970915090001@vic-ca1-11.ix.netcom.com>, LTaylor7@ix.netcom.com (Daniel L. Taylor) writes: In article <SCOTT.97Sep9005833@slave.doubleu.com>, scott@doubleu.com (Scott Hess) wrote: > Look ahead a couple years, though. Will things move that quickly > over 100Mb/s _switched_ Ethernet? How about Gigabit ethernet? > How about Gigabit switched ethernet? How about 155Mbit or higher > ATM LANs? Of course NCs aren't perfect for _today's_ networking > infrastructure - but there's something to be said for being ready > with the perfect solution when tomorrow arrives. Perfect solutions require perfect business decisions and perfect marketing and perfect PR. Neither of which Apple is demonstrating right now. Unfortunately true. On the other hand, I would rather Apple "went out" with an interesting potential solution to tomorrow's (perceived) problems than with a boring non-solution to yesterday's problems. If Apple has to become Microsoft to survive, I don't really want to hang around :-). Later, -- scott hess <scott@doubleu.com> (606) 578-0412 http://www.doubleu.com/ <Favorite unused computer book title: The Demystified Idiots Guide to the Zen of Dummies in a Nutshell in Seven Days, Unleashed>
From: shan0029@gold.tc.umn.edu (Sharad J Shanbhag) Newsgroups: comp.sys.next.programmer Subject: DO Programming Question Date: 9 Sep 1997 17:34:26 -0500 Organization: University of Minnesota Sender: shan0029@gold.tc.umn.edu Message-ID: <5v4iti$doc@gold.tc.umn.edu> Summary: Distributed Objects, interface from command-line to Workspace app Keywords: NeXT, DO, PDO, AppKit The problem: I have a command line program which performs some computations and manipulations on a binary data file. I would like to be able to view some of these data in a graph and select certain points to be marked. Ideally, these data (organized into objects) could be passed to a graphing application and viewed and marked. Sounds like a job for NeXT's Distributed Objects, no? The Trouble(s): 1) NeXT's documentation (which came with NS 3.3) is not very helpful. The articles on NeXTanswers have helped some, but not much. Are there any better examples / docs out there? 2) I can have the data objects served by the command-line program, but once the server is sent the -run message, it blocks, never to return. However, the command-line program needs to do some things with the modified objects before exiting. How can I return control gracefully to the command-line program once the graphing app is done? One solution is to avoid DO completely and simply archive the objects to a file via an NXTypedStream, unarchive them in the graphing application, graph and indicate changes, re-archive them, and unarchive in the command-line program. This seems to be unnecessarily complicated. Any ideas, suggestions, guidance is most welcomed! Cheers, Sharad Shanbhag --- Department of Neurosurgery & Grad. Program in Neuroscience University of Minnesota shanbhag@neuro.med.umn.edu shan0029@gold.tc.umn.edu
From: Arnaud Debayeux <debayeux@easynet.fr.NO_SPAM> Newsgroups: comp.sys.next.programmer Subject: Re: Custom message dispatching ? Date: Wed, 10 Sep 1997 00:47:18 +0200 Organization: Systems Group, Easynet Ltd. Message-ID: <3415D1F5.2A7102AA@easynet.fr.NO_SPAM> References: <34153CB3.FB6BA967@easynet.fr.NO_SPAM> <5v46er$rru$1@news.apple.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, I'll try to clarify a bit. In SmallTalk there is the "dependant" mechanism. An object maintains a list of objects called dependants, and notify them by message when necessary. This notification must be explicitely coded, it's not an external mechanism added to the message dispatch, it's internal to the object. The relationships dependee-dependent should be established externally to the objects, and also dynamically. This does not break the encapsulation OO rule. It's an event oriented framework that could be applied to any object of any application, not even designed in this spirit. For exemple when a method SET_X(VAL) is invoked to a MOUSE object, a notification is send to another object CURSOR, or something else if the user modified it. It's like the connected fields in Cosmo 3D/OpenGL Optimizer, and in active OODBs with ECA (Event Condition Action) rules. These rules could be specified for some classes, or methods or particular instances. In SOM/DSOM one can implement a method lookup/dispatch for each class. Is this a good pattern ? Does any high level mechanism allow this, efficiently ? (the cost must be null for not connected objects). Is there any active OODB implemented with ObjC ? Is it used by debuggers ? Is the Objective-C runtime implementation documented (in particular the dispatch and its optimizations, this could be implemented at this level) ? Thanks for any insight. Arnaud
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Intel CPU's (Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!!) Followup-To: comp.sys.next.advocacy Date: 9 Sep 1997 23:52:04 GMT Organization: Apple Computer, Inc. Message-ID: <5v4nf4$t8m$1@news.apple.com> References: <341168EE.61A1@ix.netcom.com> Benjamin Smith <benjs@ix.netcom.com> writes > Talk of using an x86 for any Apple product or MacOS machine (as > opposed to Rhapsody) truly disgusts me. The PowerPC is a superior chip, > is more modern and Apple should run its platform on this chip (it gives > MAc superior performance in graphics) and Newtons on StrongARMs. I just don't understand this attitude. Of course, I've only "seen the light" since coming to work at NeXT - the chips really don't matter, THE PLATFORM is what matters. I've run NEXTSTEP on 68040's, PowerPC's, Intel chips, PA-Risc and Sparc chips. Guess what? Some chips have better floating point performance, some have faster task switching, some are really hard to generate efficient code for...But running NEXTSTEP on all of them was the same experience, and writing applications that ran the same on all architectures was a no-brainer. Rhapsody will be the same way, smoothing over the underlying hardware. [I didn't mention the unbelievably hideous "Intel PC" hardware architecture above. I don't consider that Intel's fault - I blame IBM and Microsoft] > We expect the best out of Apple and AIM, all players, must get their act > together and not talk of killing this chip. The talk should be of > continuing to outpace the x86 and, I guess, the vapour Merced. Not this. This sounds like a goal for Motorola/IBM, not Apple. If Apple's products can be made better/faster/cheaper using Intel CPU's, then why shouldn't they? You mentioned the StrongARM above - that's a great example of using the right chip for the job. -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: "Patrick McGowan" <patmvcr@rogers.wave.ca> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software Subject: HELP FDISK disaster!!!!!!!!!! Date: 10 Sep 1997 01:48:37 GMT Organization: INCO Exploration Message-ID: <01bcbd8c$67a88620$7c327118@patmvcr.ns.bc.rogers.wave.ca> I was trying to load NextStep3.3 User in a free block on my 2.1G hard drive. The block was sandwiched between the primary DOS partition and the extended DOS partition. When prompted where to install NextStep, there was no option to leave the extended DOS partion alone and just load NS into the free space...it saw everything beyond the primary DOS partition as free space. Instead of just stopping there, i selected the option of leaving some of the "free"space as NextStep the rest for DOS......in the hope that the extended DOS partions might still be there after the install.....well I got cold feet and ditched the install, but when I came back out into the DOD/Windows95 world the extended partions were no longer receognized......ugh Are these partitions lost forever? Is there any way to recover them? I had backed up most of the files a few weeks ago, but a few important files were not. I have not reformatted those sectors of the drive or done anything else since......except look at them in Partition Magic, an interactive partitioning application in DOS. Can I use something from Norton Disk Doctor to find the beginning of my old partitions?...I had a lot af spare space between the original "free" space and my important DOD partitions...it was set aside as swap space for Windows 95. HELP!!!!!
From: greg@cnsii.com (Greg Shorts) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 21:19:39 -0800 Organization: GS Interactive Message-ID: <greg-0909972119390001@dt1h6n37.san.rr.com> References: <5v1ehb$7fk$1@news.fas.harvard.edu> <B039C792-24F3D@206.165.44.76> <5v280i$ks8@sjx-ixn5.ix.netcom.com> > >In either case, prediction isn't really an issue, apart from predicting >if anyone will want a particular NC model at all. Or want an NC at all. There really aren't a lot of companies lining up to purchase these as it is. I believe the NC is going to be a dismal failure as it is being driven almost exclusively by the IS departments as an advantage to them. Concidering that IS is a SUPPORT function in the vast majority of corporations they will have to convince the end users that there is value to these machines. I know many people will say that the users will use whatever they are given, but in this age of the knowledge worker, the idea of shoving a restrictive system down the throats is not going to make the company very successful.
From: ltest@rte9-sun_5.5.1 (news_check.py) Newsgroups: comp.soft-sys.nextstep,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,pgh.next-users,tw.bbs.comp.os.nextstep Subject: Re: Tar.gz : how can I expand this files ? Date: Wed, 10 Sep 1997 13:59:54 GMT Organization: AT&T WorldNet Load Test Distribution: inet Message-ID: <5v69df$491@lztnsc06.att.com> References: <01bcb3d8$02388e60$050e03c3@worldnet.sct.fr> In article <5ucl5k$5g4$1@xeres.hsh.stusta.mhn.de> scholz@leo.org (Bernhard Scholz) wrote: > "Laurent" <solfra@worldnet.fr> wrote: > >Hello next-web afficionados, > > "Laurent" <solfra@worldnet.fr> wrote: >Hello next-web afficionados, > > >I don't know how I can expand the files with the extension b.tar.gz with >Openstep for Mach 4.1. > The ending .b. indicates it's a binary and is of no further use. tar indicates 'TAR' archives and gz indicates 'GnuZIP. You can use gunzip and tar -x together or use gnutar -zxf <file> Greetings, Bernhard. -- Bernhard Scholz http://www.leo.org/~scholz/ Peanuts FTP Admin http://peanuts.leo.org/ scholz@leo.org, (StuSta ONLY: boerny@xenia.hsh.stusta.mhn.de)
From: ltest@rte9-sun_5.5.1 (news_check.py) Newsgroups: comp.soft-sys.nextstep,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,pgh.next-users,tw.bbs.comp.os.nextstep Subject: Re: Tar.gz : how can I expand this files ? Date: Wed, 10 Sep 1997 13:59:59 GMT Organization: AT&T WorldNet Load Test Distribution: inet Message-ID: <5v69dl$54v@lztnsc06.att.com> References: <01bcb3d8$02388e60$050e03c3@worldnet.sct.fr> In article <5ucl5k$5g4$1@xeres.hsh.stusta.mhn.de> scholz@leo.org (Bernhard Scholz) wrote: > "Laurent" <solfra@worldnet.fr> wrote: > >Hello next-web afficionados, > > "Laurent" <solfra@worldnet.fr> wrote: >Hello next-web afficionados, > > >I don't know how I can expand the files with the extension b.tar.gz with >Openstep for Mach 4.1. > The ending .b. indicates it's a binary and is of no further use. tar indicates 'TAR' archives and gz indicates 'GnuZIP. You can use gunzip and tar -x together or use gnutar -zxf <file> Greetings, Bernhard. -- Bernhard Scholz http://www.leo.org/~scholz/ Peanuts FTP Admin http://peanuts.leo.org/ scholz@leo.org, (StuSta ONLY: boerny@xenia.hsh.stusta.mhn.de)
From: ltest@rte9-sun_5.5.1 (news_check.py) Newsgroups: comp.sys.next.programmer,comp.sys.next.bugs Subject: Re: Setting nil objects as values for UI elements...wierd. Date: Wed, 10 Sep 1997 14:01:42 GMT Organization: AT&T WorldNet Load Test Message-ID: <5v69gs$3of@lztnsc06.att.com> References: <5ui9k7$io4$2@concorde.ctp.com> In article <5uj20i$cr9@sjx-ixn9.ix.netcom.com> mtrombin@ix.netcom.com (Mark Trombino) wrote: > In <5ui9k7$io4$2@concorde.ctp.com> Thomas Engel wrote: > > is it just me or are there other people who consider this as a bug: > > In <5ui9k7$io4$2@concorde.ctp.com> Thomas Engel wrote: > is it just me or are there other people who consider this as a bug: > > [myTextField setStringValue:nil] > > results in an exception...but for a textField I think it might make sense to > resert the content to the state of a freshly created textField. (I don't knwo > from the top of my head...but I think that in this case stringValue really > returns a "nil" which in this case is a value you don't expect since you > can't set it...) > > But: > > [myColorWell setColor:nil] > > works like a charm which IMHO makes absolutly no sense since the color well > will display a colorBackgroundGray "color" (or some random color) and lure > you to believe that the well contains some useful information. > It also acts a little unpredictalble...since sometimes you get some funky > color (say Pantone blue)...but when you drag the swatch it is a black swatch. I think that [myTextField setStringValue:@""] makes more sense than [myTextField setStringValue:nil] but it would be nice if setting a string value to nil had the same effect. I agree with you about the color wells though.... -- |-------------------------------------------------------------------------| | Mark Trombino | J A M S o f t | | mtrombin@ix.netcom.com | Audio DSP Tools for Openstep & Rhapsody | |-------------------------------------------------------------------------|
From: rainer@mathematik.uni-wuerzburg.de (Rainer =?ISO-8859-1?Q?Frohnh=F6fer)?= Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 9 Sep 1997 19:31:13 GMT Organization: University of =?ISO-8859-1?Q?W=FCrzburg,?= Germany Message-ID: <5v4861$1rs@lobotomy.urz.uni-wuerzburg.de> References: <5v1ehb$7fk$1@news.fas.harvard.edu> <B039C792-24F3D@206.165.44.76> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: english@primenet.com In <B039C792-24F3D@206.165.44.76> "Lawson English" wrote: > Ishir Bhan <ibhan@digitas.harvard.edu> said: > > > Why do you think that Jobs plan is to make NC's instead of Rhapsody > boxes? > > Let's face it: the MacOS is really at its limit. Rhapsody for PC's and > > either a modified Rhapsody for or Newton OS for NC's is probably where > > Apple is headed. > > > > What is your problem with this? > > The margin for NC's is likely to be VERY slight. APple has had trouble in > forcasting and meeting the demand for its far more profitable boxes (that's > why the claim that Power was stealing from them makes NO sense -even with > 50% margins, if you can't ship to your customers on time, a 50% margin on > zero is still zero% profit) and it seems foolish to dedicate manufacturing > capabilities that could go to the high-end to the lowest of the low-end. I don't think the people that came up with the NC concept wanted to make lots of money from them. It's a Sun idea and Sun sells servers. I guess the idea was to shell out cheap NCs and then charge *real* money for the big iron to supply sufficient horsepower; and the software, of course. Sun is fighting MS's grip on the desktop in the business world. A home NC wouldn't make too much sense until the bandwith is there. Even then I doubt it's going to be a Sun box, rather a Sony TV with built in Java & MPEG or whatever chips. > And if Apple expects clone-makers to line up for NC designs, who would now > trust them to keep to agreements and step-forward? Good point. ------------------------------------- "Um Energie zu sparen, wird das Licht am Ende des Tunnels vorlaeufig abgeschaltet." rainer@mathematik.uni-wuerzburg.de (public key avaible at any key server near you ...)
Newsgroups: comp.soft-sys.nextstep,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,pgh.next-users,tw.bbs.comp.os.nextstep From: bofh@keltia.freenix.fr (Ollivier Robert) Sender: ltest@rte9-sun_5.5.1 (news_check.py) Message-ID: <cancel.5v6a51$ga3@lztnsc06.att.com> Subject: cmsg cancel <5v6a51$ga3@lztnsc06.att.com> ignore Control: cancel <5v6a51$ga3@lztnsc06.att.com> Organization: Usenet Canal Historique Date: Wed, 10 Sep 1997 14:12:27 GMT Distribution: inet ECP/EMP aka SPAM or pyramidal scheme (MMF) cancelled by bofh@keltia.freenix.fr. It may also be an image too small for newsbot to be activated. See report in news.admin.net-abuse.bulletins. Date: Wed Sep 10 17:42:50 1997 Original subject was: Re: Tar.gz : how can I expand this files ?
Newsgroups: comp.soft-sys.nextstep,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,pgh.next-users,tw.bbs.comp.os.nextstep From: bofh@keltia.freenix.fr (Ollivier Robert) Sender: ltest@rte9-sun_5.5.1 (news_check.py) Message-ID: <cancel.5v69df$491@lztnsc06.att.com> Subject: cmsg cancel <5v69df$491@lztnsc06.att.com> ignore Control: cancel <5v69df$491@lztnsc06.att.com> Organization: Usenet Canal Historique Date: Wed, 10 Sep 1997 13:59:54 GMT Distribution: inet ECP/EMP aka SPAM or pyramidal scheme (MMF) cancelled by bofh@keltia.freenix.fr. It may also be an image too small for newsbot to be activated. See report in news.admin.net-abuse.bulletins. Date: Wed Sep 10 17:42:41 1997 Original subject was: Re: Tar.gz : how can I expand this files ?
Newsgroups: comp.soft-sys.nextstep,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,pgh.next-users,tw.bbs.comp.os.nextstep From: bofh@keltia.freenix.fr (Ollivier Robert) Sender: ltest@rte9-sun_5.5.1 (news_check.py) Message-ID: <cancel.5v69dl$54v@lztnsc06.att.com> Subject: cmsg cancel <5v69dl$54v@lztnsc06.att.com> ignore Control: cancel <5v69dl$54v@lztnsc06.att.com> Organization: Usenet Canal Historique Date: Wed, 10 Sep 1997 13:59:59 GMT Distribution: inet ECP/EMP aka SPAM or pyramidal scheme (MMF) cancelled by bofh@keltia.freenix.fr. It may also be an image too small for newsbot to be activated. See report in news.admin.net-abuse.bulletins. Date: Wed Sep 10 17:42:44 1997 Original subject was: Re: Tar.gz : how can I expand this files ?
Newsgroups: comp.sys.next.advocacy,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,de.comp.sys.next From: bofh@keltia.freenix.fr (Ollivier Robert) Sender: ltest@rte9-sun_5.5.1 (news_check.py) Message-ID: <cancel.5v6a3f$fdt@lztnsc06.att.com> Subject: cmsg cancel <5v6a3f$fdt@lztnsc06.att.com> ignore Control: cancel <5v6a3f$fdt@lztnsc06.att.com> Organization: Usenet Canal Historique Date: Wed, 10 Sep 1997 14:11:38 GMT ECP/EMP aka SPAM or pyramidal scheme (MMF) cancelled by bofh@keltia.freenix.fr. It may also be an image too small for newsbot to be activated. See report in news.admin.net-abuse.bulletins. Date: Wed Sep 10 17:47:16 1997 Original subject was: Re: S: PGP for NS 3.3 Intel
Newsgroups: comp.sys.next.programmer,comp.sys.next.bugs From: bofh@keltia.freenix.fr (Ollivier Robert) Sender: ltest@rte9-sun_5.5.1 (news_check.py) Message-ID: <cancel.5v69gs$3of@lztnsc06.att.com> Subject: cmsg cancel <5v69gs$3of@lztnsc06.att.com> ignore Control: cancel <5v69gs$3of@lztnsc06.att.com> Organization: Usenet Canal Historique Date: Wed, 10 Sep 1997 14:01:42 GMT ECP/EMP aka SPAM or pyramidal scheme (MMF) cancelled by bofh@keltia.freenix.fr. It may also be an image too small for newsbot to be activated. See report in news.admin.net-abuse.bulletins. Date: Wed Sep 10 17:47:20 1997 Original subject was: Re: Setting nil objects as values for UI elements...wierd.
From: ltest@rte9-sun_5.5.1 (news_check.py) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Wed, 10 Sep 1997 14:00:52 GMT Organization: AT&T WorldNet Load Test Message-ID: <5v69f9$5an@lztnsc06.att.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <acurylo-0509971010020001@p2-29.van.tvs.net> <5ur2uv$r3s@crcnis3.unl.edu> <5uujej$2to$1@nntp.ucs.ubc.ca> Delivery-Date: Wed, 10 Sep 1997 07:03:44 -0700 Return-Path: lztnsc06!newsadm@uunet.uu.net for <gherbert@crl.com>; Wed, 10 Sep 1997 07:00:46 -0700 (PDT) env-from (lztnsc06!newsadm@uunet.uu.net) id QQdgke04757; Wed, 10 Sep 1997 10:05:39 -0400 (EDT) To: sci-space-tech@uunet.uu.net In article <5v2970$719@peabody.colorado.edu> fcrary@rintintin.Colorado.EDU (Frank Crary) wrote: > > In article <5uujej$2to$1@nntp.ucs.ubc.ca>, > Tim Shoppa <shoppa@alph02.triumf.ca> wrote: In article <5uujej$2to$1@nntp.ucs.ubc.ca>, Tim Shoppa <shoppa@alph02.triumf.ca> wrote: >>: Not to mention the Mars Rover. (PPC -- 100% of the >>: Martian computing installed base! There's an ad for you huh?) >>Not to rain on your extraterrestial parade or anything, but >>doesn't the Mars Rover use a version of the Intel 8085? >Yes, a 80C85 is in the Rover. The lander has a IBM RAD6000, which >is vaguely related to the PowerPC (but it is *not* a PowerPC). Two comments. First, I though it was a RAD4000, which would be a radiation hardened version of the R4000 chip. Second, last time I checked, the PowerPC was a computer (or series of computers), not a chip. So the connection to PowerPCs is simply that they use the a similar processor. (Nor are PowerPC the only computers that use this family of chips: SGI uses them extensively; the Indigo II on my desk at work uses a R4400 processor.) I think it's important to avoid confusing chips with computers. Frank Crary CU Boulder
From: tom@basil.icce.rug.dev.null.nl (Tom Hageman) Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Exception processing in Objective C? Date: 10 Sep 1997 17:42:13 GMT Organization: Warty Wolfs Sender: news@basil.icce.rug.nl (NEWS pusher) Message-ID: <EG9Ds3.76A@basil.icce.rug.nl> References: <340C6C07.5727@win.bright.net> <5v1f07$loh@shelob.afs.com> <873800789.32459@dejanews.com> David Stes <stes@can.nl> wrote: > Rich Schroedel <richs@win.bright.net> writes > > I'm in the process of rewriting some C++ code so that it can eventually > > be converted to Objective C. I'm stumped at how to handle exception > > processing. How do Objective C programmers handle the sort of situations > > that C++ programmers handle with exception processing? Any help out > > there? > > I think this can best be done by using Objective-C Blocks. Caveat: Blocks are a unique feature of David's Portable Objective C compiler. The other Objective-C implementations (GNU, NeXT/Apple, Stepstone) do not support it (yet?). Check the Objective-C FAQ at ftp://ftp.ics.ele.tue.nl/pub/objc/answers (question 17) to find out about the differences between the various compilers and runtimes. -- __/__/__/__/ Tom Hageman <tom@basil.icce.rug.dev.null.nl> [NeXTmail/Mime OK] __/ __/_/ IC Group <tom@icgned.dev.null.nl> (work) __/__/__/ <<SPAMBLOCK: remove .dev.null to reply>> __/ _/_/ Confused? You won't be after the NeXT episode.
From: bdm@vayu.psych.nyu.edu Newsgroups: comp.sys.next.programmer Subject: g77 in OS4.2 Date: Wed, 10 Sep 1997 14:19:24 GMT Organization: New York University Message-ID: <970910101924.401AAG1G.bdm@vayu> Mime-Version: 1.0 (Generated by Eloquent) Content-Type: text/plain; charset=US-ASCII Does anyone have src or binaries for g77 that will work under OpenStep 4.2? Thanks in advance, Brian Brian McElree Program in Cognition & Perception Department of Psychology New York University <bdm@xp.psych.nyu.edu> NeXT Mail <bdm@vayu.psych.nyu.edu>
From: gutier@unixg.ubc.ca (Gerald Gutierrez) Newsgroups: comp.sys.next.programmer Subject: NeXT palette manipulations ... Date: 10 Sep 1997 19:19:28 GMT Organization: University of British Columbia, Vancouver, B.C., Canada Message-ID: <5v6rs0$mn0$1@nntp.ucs.ubc.ca> I have long been intrigued by the way NeXTSTEP manages an 8 bit palette. With 256 colors, it is able to display all sorts of graphics with different color makeups, and can even do transparency speedily, and without messing up the colors of other objects on the screen. Does anyone know HOW NeXTSTEP manages its 8 bit palette ? It seems quite magical to me, and I'm interested in the tricks and techniques that it employs. Thanks.
From: nothanks@ix.netcom.com (John Doherty) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Tue, 09 Sep 1997 21:01:24 -0500 Organization: (none) Message-ID: <nothanks-0909972101250001@aus-tx22-14.ix.netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <acurylo-0509971010020001@p2-29.van.tvs.net> <5ur2uv$r3s@crcnis3.unl.edu> <5uujej$2to$1@nntp.ucs.ubc.ca> <5v2970$719@peabody.colorado.edu> Delivery-Date: Tue, 09 Sep 1997 19:08:24 -0700 Return-Path: news@sjx-ixn4.ix.netcom.com for <gherbert@crl.com>; Tue, 9 Sep 1997 18:56:22 -0700 (PDT) env-from (news@sjx-ixn4.ix.netcom.com) id QQdgii26472; Tue, 9 Sep 1997 22:00:51 -0400 (EDT) id TAA26533; Tue, 9 Sep 1997 19:00:56 -0700 (PDT) To: sci-space-tech@uunet.uu.net In article <5v2970$719@peabody.colorado.edu>, fcrary@rintintin.Colorado.EDU (Frank Crary) wrote: | Second, last time I checked, the PowerPC was a computer | (or series of computers), not a chip. You need to check again. PowerPC is a family of microprocessors designed and manufactured by Motorola and IBM. Members of the family are the PPC603, PPC603e, PPC604, PPC604e, etc. Various computer systems are built using PowerPC microprocessors, including PowerMacs and compatibles from Apple and others, and systems built by IBM and Motorola (e.g., the IBM RS/6000 which uses multiple PPC processors in an SMP configuration or a single POWER2 processor). Deep Blue, the IBM machine that beat Kasparov at chess, is an RS/6000 with 64 (I believe) PowerPC processors. | (Nor are PowerPC the only computers that use this family of chips: SGI uses | them extensively; the Indigo II on my desk at work uses a R4400 processor.) An R4400 is another microprocessor, but it is not a PowerPC. It's designed and built by MIPS (now a subsidiary of SGI, I believe). To recap: a PowerPC is a microprocessor, not a computer. Your SGI computer has a microprocessor, but that microprocessor is not a PowerPC. | I think it's important to avoid confusing chips with computers. I agree. ;)
From: bettis@inetnebr.com (Mr. Jeremy Bettis) Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Linking error on bundle Date: 10 Sep 1997 14:58:48 -0500 Organization: Internet Nebraska Message-ID: <5v6u5o$5cv$1@falcon.inetnebr.com> References: <34142B50.41C6@ina.fr> NNTP-Posting-User: bettis Philippe Converset <plc@ina.fr> writes: >Here is a very simple example which produces a link error in NT OpenStep >Enterprise 4.2 : >The result of compilation is : >ClassB.o : error LNK2001 unresolved external symbol >.objc_class_name_ClassA >It works fine under Mach but not under NT. >Did I forget a compiler or linker flag? >Is NT not supporting loadable bundles? >Is it a Project Builder bug? >Any idea????? Add OTHER_LDFLAGS += -Xlinker -force to Makefile.preamble (or Makefile.preamble-winnt as I like to do, then add -include Makefile.preamble-$(PLATFORM_OS) to the end of Makefile.preamble) It is a bug in the Project builder makefiles. Has been there since Pre-release 1. I don't think they ever fix bugs, just cover them up with new bugs. your applications. Transparency, etc, is also handled by the WindowServer. All of this is possible, because drawing is never done 'directly to the screen', but always through the WindowServer. The WindowServer is your friend :) > >Thanks. > // Christian Brunschen
From: Henry Spencer <henry@zoo.toronto.edu> Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Wed, 10 Sep 1997 02:47:10 GMT Organization: SP Systems, Toronto Message-ID: <EG9uEM.K69%spenford@zoo.toronto.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <5ur2uv$r3s@crcnis3.unl.edu> <5uujej$2to$1@nntp.ucs.ubc.ca> <5v2970$719@peabody.colorado.edu> In article <5v2970$719@peabody.colorado.edu>, Frank Crary <fcrary@rintintin.Colorado.EDU> wrote: >>Yes, a 80C85 is in the Rover. The lander has a IBM RAD6000, which >>is vaguely related to the PowerPC (but it is *not* a PowerPC). > >Two comments. First, I though it was a RAD4000, which would be a >radiation hardened version of the R4000 chip... No, the R4000 is a MIPS design, which is unrelated to the IBM family tree of CPU designs that includes the RS/6000 and the PowerPC chips. (There is some common heritage from the early RISC concepts, but the MIPS and IBM processors have no real common ancestors.) -- The operating systems of the 1950s will be out | Henry Spencer next year from Microsoft. -- Mark Weiser | henry@zoo.toronto.edu
From: ga@ed4u.com (G. Apple) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: CPUs in Space (Was Re: No more PowerPCs... Date: 10 Sep 1997 17:18:35 GMT Organization: Advanced Communications Engineering, Inc. Message-ID: <5v6kpb$mre$1@gte2.gte.net> References: <5uujej$2to$1@nntp.ucs.ubc.ca> <B039C5D9-1E794@206.165.44.76> <341556f5.560572@news.gatech.edu> Delivery-Date: Wed, 10 Sep 1997 10:17:01 -0700 Return-Path: news@roadrunner.gte.net for <gherbert@crl.com>; Wed, 10 Sep 1997 10:11:14 -0700 (PDT) env-from (news@roadrunner.gte.net) by gte2.gte.net (8.8.7/8.8.7) id MAA20789; Wed, 10 Sep 1997 12:18:35 -0500 (CDT) To: sci-space-tech@ncren.net In article <341556f5.560572@news.gatech.edu>, bmeadows@acl1.physics.gatech.edu (Brian K. Meadows) wrote: > There are several know "process receipts" for CMOS, BJT, and Bi-CMOS > transistors that include hardening against radiation. Rad-hardened > chips are made in quantity for the military. Engineers for Pathfinder > could always use shielding, but I doubt the designers wanted to pay > the weight penalty. > Most radiation hardening involves things like series resistance in paths that that are subject to short circuits. "Semiconductors" are primarily insulators. Radiation has the same effect on a semiconductor as you see in a cloud chamber, i.e., ionization trails galore. These ions (holes and electrons in this case) turn that sucker into a full conductor, thus shorting many paths between power and ground, potentially causing local overheating and failure of the chip. Lesser energetic (soft) radiation may simply flip a few bits by discharging some charge wells. Military satellites generally have extensive radiation hardening and shielding for obvious reasons. From what little I know about pathfinder (mostly from a small meeting of the LA Satellite Professionals group last month at JPL) and from my own involvement as a systems engineer with many other NASA programs, this mission was done on the cheap. I doubt that much shielding protection was provided. -- G. Gordon Apple, PhD The Ed4U Project Advanced Communications Engineering, Inc. Redondo Beach, CA ga@ed4u.com
From: don_arb@wolfenet.com (Don Arbow) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Wed, 10 Sep 1997 12:14:04 -0700 Organization: None Message-ID: <don_arb-1009971214040001@sea-ts5-p12.wolfenet.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0709971238090001@vic-ca1-13.ix.netcom.com> <5v0h4t$1ir$1@news.lth.se> <LTaylor7-0809971150040001@vic-ca1-12.ix.netcom.com> <5v33e2$3pk$1@news.lth.se> <LTaylor7-0909970925550001@vic-ca1-11.ix.netcom.com> In article <LTaylor7-0909970925550001@vic-ca1-11.ix.netcom.com>, LTaylor7@ix.netcom.com (Daniel L. Taylor) wrote: : : > (we have : > seen three cloners gain MacOS 8 licenses, and now there are these : > rumors about continued talk between Apple and Motorola); Rhapsody : > (which was a big part of Amelio's strategy) is coming along better : > than expected. : : Really? Was the DR not to be out in the summer? Is it just me, or did : it slip? : The Rhapsody DR1 schedule has not slipped. As I recall from the announcement at WWDC last May, DR1 was promised for August/September. According to today's MacOSRumors, DR1 will go golden in the next 10 days. Rhapsody development overall has been on time and on schedule since porting started last February. Yellow Box development is reported to be far ahead of schedule. Don -- Don Arbow, Partner, CTO EveryDay Objects, Inc. don_arb@wolfenet.com <-- remove underscore to reply http://www.edo-inc.com
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT palette manipulations ... Date: 10 Sep 1997 20:59:01 GMT Organization: Apple Computer, Inc. Message-ID: <5v71ml$lsu$1@news.apple.com> References: <5v6rs0$mn0$1@nntp.ucs.ubc.ca> Gerald Gutierrez writes > > I have long been intrigued by the way NeXTSTEP manages an 8 bit palette. > With 256 colors, it is able to display all sorts of graphics with > different color makeups, and can even do transparency speedily, and > without messing up the colors of other objects on the screen. > > Does anyone know HOW NeXTSTEP manages its 8 bit palette ? It seems quite > magical to me, and I'm interested in the tricks and techniques that it > employs. > > Thanks. It's all part of the magic of PostScript :-) The video palette gets set once during startup to a "reasonable" set of colors, and never gets changed thereafter. The WindowServer simply(!) dithers colors down from their internal representation to whatever the screen is. This is a real 80/20 sort of solution. There are cases where it doesn't work very well, but most of the time, it works great. Contrast this to the palette-swapping "solution" utilized by most windowing systems. You can make *any one* window look great , but the rest of the screen looks like crap... -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: Custom message dispatching ? Date: 10 Sep 1997 20:52:22 GMT Organization: Apple Computer, Inc. Message-ID: <5v71a6$120o$1@news.apple.com> References: <3415D1F5.2A7102AA@easynet.fr.NO_SPAM> Arnaud Debayeux <debayeux@easynet.fr.NO_SPAM> writes > Hi, > > I'll try to clarify a bit. > > In SmallTalk there is the "dependant" mechanism. An object maintains a > list of objects called dependants, and notify them by message when > necessary. This notification must be explicitely coded, it's not an > external mechanism added to the message dispatch, it's internal to the > object. Okay, I was right with you up to that last sentence. So each object has an instance variable containing a list of "dependent" objects? That's easy to do in ObjC. > The relationships dependee-dependent should be established externally to > the objects, and also dynamically. This does not break the encapsulation > OO rule. It's an event oriented framework that could be applied to any > object of any application, not even designed in this spirit. Wait, "externally to the object"? I don't get it. Do you mean that the relationship is stored external to the object, or just that they can be managed externally? The second case is trivial again: just add "addDependant:" and "removeDependant:" methods to the class. However, adding this behavior to classes "not even designed in this spirit" seems like a challenge. You can't (easily) add instance variables to someone else's classes. > For exemple when a method SET_X(VAL) is invoked to a MOUSE object, a > notification is send to another object CURSOR, or something else if the > user modified it. There's a notification mechanism in OPENSTEP which can be used to do this, but your class has to have code in it to send the notifications. It doesn't need to worry about who's listening for tham, though. If you want a mechanism where, external to the implementation of either of the classes, you can say: "when this instance of the Mouse class receives the set_x message, send a mouse_moved message to this instance of the Cursor class", that's gonna be a challenge. > It's like the connected fields in Cosmo 3D/OpenGL Optimizer, and in > active OODBs with ECA (Event Condition Action) rules. These rules could > be specified for some classes, or methods or particular instances. In > SOM/DSOM one can implement a method lookup/dispatch for each class. Unfortunately, I'm not familiar with any of these examples :-( > Is this a good pattern ? Assuming that I've finally interpreted it correctly, it's certainly *interesting*. I don't know that I like the idea of keeping relationship information separate from the objects involved. It seems like you'e moving some information out of the class that it really needs to know about. > Does any high level mechanism allow this, efficiently ? (the cost must > be null for not connected objects). Not that I know of. All the messaging in Objective-C goes through a central point, a function called objc_msgSend(). You could modify that function to support this functionality, I guess... > Is there any active OODB implemented with ObjC ? Don't know. > Is it used by debuggers ? No. Most people use gdb (GNU debugger) to debug their ObjC code. It's very much a C debugger, not at all like the Smalltalk debugger. > Is the Objective-C runtime implementation documented (in particular the > dispatch and its optimizations, this could be implemented at this level) It's "kind of" documented. If you use the GNU runtime, you can of course get the source code. If you're using the NeXT/Apple runtime, there's no comprehensive reference for it (yet). -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: hhoff@ragnarok.en.REMOVETHIS.eunet.de (Holger Hoffstaette) Newsgroups: comp.sys.next.programmer Subject: Re: Custom message dispatching ? Date: 10 Sep 1997 22:32:46 GMT Organization: N.eXTers W.ith A.ttidude Message-ID: <5v776e$djh@ragnarok.en.eunet.de> References: <34153CB3.FB6BA967@easynet.fr.NO_SPAM> <5v46er$rru$1@news.apple.com> <3415D1F5.2A7102AA@easynet.fr.NO_SPAM> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Arnaud Debayeux wrote: > In SmallTalk there is the "dependant" mechanism. An object maintains a > list of objects called dependants, and notify them by message when > necessary. This notification must be explicitely coded, it's not an > external mechanism added to the message dispatch, it's internal to > the object. Something similar exists in OPENSTEP - there's no need for you to modify the runtime or the message dispatch mechanism in any way. The mechanism is called 'notification', implemented by the class NSNotification (a broadcast message). This object is dispatched by an NSNotificationCenter, which registers and unregisters objects for the dispatch. If any object 'posts' a notification with a certain name to a NSNotificationCenter (there's a default center, which is sufficient most of the time) the center performs a selector on any objects that are registered for the message by the particular sender. Note that NSNotification flow can be channeled in two dimensions: for one, you can control who receives the notifications. Second, it's possible for an object to listen for notifications by a specified sender only. This helps to cut down on unnecessary dispatch traffic. > The relationships dependee-dependent should be established externally to the > objects, and also dynamically. This does not break the encapsulation OO rule. NSNotificationCenter can add and remove objects ('observers') dynamically, without the observers really knowing that they've been registered. When a notification is broadcast to all its registered observers, each one will receive the registered message. > It's an event oriented framework that could be applied to any object of any > application, not even designed in this spirit. Yes. For everybody who's designing dynamic collaboration-based architectures it's a godsend. > For exemple when a method SET_X(VAL) is invoked to a MOUSE object, a > notification is send to another object CURSOR, or something else if the user > modified it. It's like the connected fields in Cosmo 3D/OpenGL Optimizer, and > in active OODBs with ECA (Event Condition Action) rules. These rules could be > specified for some classes, or methods or particular instances. See above! > In SOM/DSOM one can implement a method lookup/dispatch for each class. I don't know anything about SOM/DSOM, except that SOM is a language binding, sort of like IDL. > Is this a good pattern ? > Does any high level mechanism allow this, efficiently ? (the cost must be > null for not connected objects). Yes. If your object is not registered to receive any notifications, no dispatch will occur for that particular object. Makes sense, too. :-) > Is there any active OODB implemented with ObjC ? Not that I know of. Would be great to have..but unless major strategic directions for EOF (and/or Apple) become clear, it's probably too early to think about this. Also, note that you'll get persistent Java object stores for free with Java in Rhapsody. It would certainly be possible, with the new EOF2 architecture.. (everybody: if you want to see an OODBMS for Rhapsody, send email with suggestions or comments to holger"at"object-factory.com - that's my address at work. They pay me to do cool things. :) > Is it used by debuggers ? I guess the 4.2 PB/gdb combination makes use of this; maybe via the DevKit framework (just guessing). Notifications can be sent across thread, task and machine (!) boundaries via PDO, so it's much more useful than the stock messaging in Smalltalk, with some gross CORBA overload or DCOM junk hacked in, just because you want to message a remote process (asynchronously, if you wish). > Is the Objective-C runtime implementation documented (in particular the > dispatch and its optimizations, this could be implemented at this level) ? The runtime has hooks, but I'm not sure you want to deal with it unless you really know what you're doing. The high-level construct with NSNotification, NSNotificationCenter and its associated classes is good enough, unless you want to broadcast mouseclicks to hundreds of objects several times a second. If performance is really an issue, an alternative way would be to have an NSArray as ivar, and then have all the dependents in the array messaged via NSArray's built-in method [dependentsArray makeObjectsPerformSelector: @selector(foo)]; I'm not sure if that's really significantly faster..maybe if you directly call the IMP pointers, like I usually do.. =:) hope this gives you some ideas Holger PS: Mark B.: when do we get [aClass allInstances]? Should be easy. :) -- holger"at"ragnarok.en.eunet.de (NeXTMail, MIME) 'Where I come from it's important to develop stuff that doesn't suck.' -- David Young <dwy"at"ace.net>
From: sef@kithrup.com Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: cmsg cancel <5v69f9$5an@lztnsc06.att.com> Date: 10 Sep 1997 23:57:14 GMT Control: cancel <5v69f9$5an@lztnsc06.att.com> Message-ID: <cancel.5v69f9$5an@lztnsc06.att.com> Sender: ltest@rte9-sun_5.5.1 (news_check.py) Spam cancelled by sef@kithrup.com
From: Nathan Keir Edel <edel@best.com_SPAMBLOCK> Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 10 Sep 1997 20:09:43 GMT Organization: Forte Systems, Inc. Message-ID: <5v6uq7$5na$2@nntp2.ba.best.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <acurylo-0509971010020001@p2-29.van.tvs.net> <5ur2uv$r3s@crcnis3.unl.edu> <5uujej$2to$1@nntp.ucs.ubc.ca> <5v2970$719@peabody.colorado.edu> Delivery-Date: Wed, 10 Sep 1997 13:24:52 -0700 Return-Path: news@nntp2.ba.best.com for <gherbert@crl.com>; Wed, 10 Sep 1997 13:05:26 -0700 (PDT) env-from (news@nntp2.ba.best.com) To: sci-space-tech@moderators.uu.net In comp.sys.mac.advocacy Frank Crary <fcrary@rintintin.Colorado.EDU> wrote: FC% I checked, the PowerPC was a computer (or series of computers), FC% not a chip. So the connection to PowerPCs is simply that they FC% use the a similar processor. (Nor are PowerPC the only computers FC% that use this family of chips: SGI uses them extensively; the FC% Indigo II on my desk at work uses a R4400 processor.) I think it's FC% important to avoid confusing chips with computers. Wrong. PowerPC is the trademark for a series of chips from Motorola and IBM, including the PowerPC 600 series used mainly in Macs and several other IBM and Motorola serieses used in RS/6000 workstations and embedded devices. The MIPS chips used by SGI machines, like the R4000 series, are unrelated except for the basic RISC concept. The PowerPC chips are descended from the Power-II chipset used in older IBM RS/6000 systems. -- Nathan Keir Edel "In this world of delusion, #6: Which side are you on? edel@best.com never turn your back on a #2: That would be telling. friend..." -- Iron Maiden -- The Prisoner
From: nurban@sps1.phys.vt.edu (Nathan Urban) Newsgroups: comp.sys.next.programmer Subject: TCPTransport example Date: 10 Sep 1997 21:01:23 -0400 Organization: Data Systems Consulting, Inc. Message-ID: <5v7ft3$p37$1@sps1.phys.vt.edu> Has anyone gotten this to work under OSfM 4.1? I have no idea what the problem is. Here's the gdb output and stack backtrace: Starting program: /usr/local/Users/nurban/Developer/TCPTransport/TCPTransport Sep 10 20:48:23 TCPTransport[3450] listening on TCP port: <TCPPort 10000, addr 1073741834, reader 0, readWriter 3> Dynamic Linkeditor at 0x12000000 offset 0x0 Executable at 0x2000 offset 0x0 /NextLibrary/Frameworks/Foundation.framework/Versions/B/Foundation at 0x18000000 offset 0x0 Inserting deferred breakpoint #1 ([NSException raise]). Breakpoint 1 at 0x1805b1f3 /NextLibrary/Frameworks/System.framework/Versions/A/System at 0x5000000 offset 0x0 Breakpoint 1, 0x1805b1f3 in -[NSException raise] () (gdb) bt #0 0x1805b1f3 in -[NSException raise] () #1 0x180396aa in +[NSException raise:format:arguments:] () #2 0x18039625 in +[NSException raise:format:] () #3 0x180aff2e in -[NSConcreteFileHandle performFunction:modes:activity:] () #4 0x18048e96 in -[NSConcreteFileHandle readInBackgroundAndNotifyForModes:] () #5 0x3e21 in -[TCPPort addConnection:toRunLoop:forMode:] (self=0xa075c, _cmd=0x180de302, conn=0xa1e7c, runLoop=0xa2ea4, mode=0x180c5fb8) at TCPPort.m:550 #6 0x1802c3bf in -[NSConnection addPortsToRunLoop:] () #7 0x1802c4ed in -[NSConnection addRunLoop:] () #8 0x1802d7d0 in -[NSConnection initWithReceivePort:sendPort:] () #9 0x180259b8 in +[NSConnection connectionWithReceivePort:sendPort:] () #10 0x5213 in main () at TCPTransport_main.m:96 #11 0x2ede in start () at crt0.c:144 Stderr has the following: Sep 10 20:56:08 TCPTransport[3459] *** Uncaught exception: <NSInternalInconsistencyException> a different activity is already in progress
Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc From: adt@netcom.com (Anthony D. Tribelli) Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Message-ID: <adtEGBp6o.It4@netcom.com> Followup-To: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Organization: Netcom On-Line Services References: <5urae7$f7c$1@attila.apana.org.au> <5utdrs$41l$1@attila.apana.org.au> <5uuqf9$4n4$1@ironhorse.plsys.co.uk> <adt-ya023080000709971749120001@news.deltanet.com> <5v0465$guj$1@attila.apana.org.au> Date: Thu, 11 Sep 1997 02:49:35 GMT Sender: adt@netcom.netcom.com Matt McLeod (mjm@attila.apana.org.au) wrote: : Tony Tribelli <adt@netcom.com> wrote: : :Matt McLeod wrote: : :> Maybe Apple just need to hire Linus Torvalds to run their OS-development. : : : :Why, his OS sales figures are far lower? :-) : : Because the free Unix people (Linux, *BSD*, etc) have been able to support : quite a few hardware platforms without the resources of a huge corporation : behind them. : : If they can do that, imagine what they could do with a little money... We don't have to imagine, we have Next, or more accurately Apple has Next. Tony -- ------------------ Tony Tribelli adtribelli@acm.org
From: 00093182@bigred.unl.edu (Josh Hesse) Newsgroups: comp.sys.next.programmer Subject: Re: NeXT palette manipulations ... Date: 11 Sep 1997 03:58:04 GMT Organization: University of Nebraska--Lincoln Message-ID: <5v7q8c$amp@crcnis3.unl.edu> References: <5v6rs0$mn0$1@nntp.ucs.ubc.ca> <5v71ml$lsu$1@news.apple.com> Mark Bessey (MaRK_BeSSeY@Apple.Com) wrote: : : It's all part of the magic of PostScript :-) The video palette gets set : once during startup to a "reasonable" set of colors, and never gets : changed thereafter. The WindowServer simply(!) dithers colors down from : their internal representation to whatever the screen is. : : This is a real 80/20 sort of solution. There are cases where it doesn't : work very well, but most of the time, it works great. Contrast this to the : palette-swapping "solution" utilized by most windowing systems. You can : make *any one* window look great , but the rest of the screen looks like : crap... Heck, I'm still amazed at how well it works with *2-bit grayscale*. I haven't done something yet that wasn't totally legible, though I'm sure it will probably happen sometime. -Josh
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer Subject: How to tell if an app is leaking memory/sucking swap (when should be idle)? Date: 11 Sep 1997 05:09:41 GMT Organization: none Message-ID: <5v7uel$fmm$1@ha1.rdc1.nj.home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I've been using a few apps which I haven't been using before. One of them seems to make the swapfile grow even when it isn't doing anything. Is there some way I can test to find which one it is? The last time I asked this, I got a lot of responses suggesting ``quit OmniWeb'' or something else like that. None of these apps is OmniWeb, or any other OmniGroup application. I'd just like a way that I, a user without source code, can run an app (perhaps over a series of days) to see if it is using more memory than it should be. This may be a simple question, but I don't know where to look. Is ``ps'' a good reference? If so, what fields are and what are not? Thanks TjL -- I'm sick of SPAM. Therefore, my email address is invalid. To email me, use one of these without spaces: luomat + next @ luomat.peak.org (for NeXT related stuff) luomat + peak @ luomat.peak.org (for PEAK related stuff)
From: pete@ohm.york.ac.uk (-bat.) Newsgroups: comp.sys.next.programmer Subject: C++ under 4.2 Date: 11 Sep 1997 05:58:21 GMT Organization: The University of York, UK Sender: pcf1@york.ac.uk Message-ID: <5v819t$kqh$1@netty.york.ac.uk> Hmmm... so, whats the cunning magic invocation you needs to get C++ working under 4.2 developer then ? I've already discovered that the compiler chnaged names to cc++, but I have yet to find where iostream.h and the rest of the header files ahve moved to - let alone tring to find out what support libraries I need to link in ! -bat. [up all night trying to solve this and very stressed]
From: arti@lava.DOTnet (Art Isbell - replace "DOT") Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Linking error on bundle Date: 11 Sep 1997 06:12:02 GMT Organization: LavaNet, Inc. Distribution: world Message-ID: <5v823i$gut@mochi.lava.net> References: <34142B50.41C6@ina.fr> <5v6u5o$5cv$1@falcon.inetnebr.com> bettis@inetnebr.com (Mr. Jeremy Bettis) wrote: > > Philippe Converset <plc@ina.fr> writes: > >Here is a very simple example which produces a link error in NT OpenStep > >Enterprise 4.2 : > > >The result of compilation is : > >ClassB.o : error LNK2001 unresolved external symbol > >.objc_class_name_ClassA > > >It works fine under Mach but not under NT. > >Did I forget a compiler or linker flag? > >Is NT not supporting loadable bundles? > >Is it a Project Builder bug? > >Any idea????? > > Add > OTHER_LDFLAGS += -Xlinker -force > to Makefile.preamble (or Makefile.preamble-winnt as I like to do, then > add -include Makefile.preamble-$(PLATFORM_OS) to the end of > Makefile.preamble) > > It is a bug in the Project builder makefiles. Has been there since > Pre-release 1. I don't think they ever fix bugs, just cover them up with new > bugs. Whoa!! Are you suggesting that -Xlinker -force or -Xlinker /FORCE or -undefined suppress or -undefined warning or whatever other variations that seem to force linking regardless of unresolved symbols be automatically included in any PB makefiles? Not in my makefiles you don't :-) All loadable bundles don't necessarily reference undefined symbols, so I don't want to risk creating a bundle that will crash an app when loaded. If a loadable bundle references symbols that are *guaranteed* to be defined in the module that's loading it, then using one of these linker flags to force linking is about all you can do with the Windows linker. But you'd better carefully inspect all unresolved symbol messages to ensure that some unintended unresolved symbols didn't sneak in as well. If a symbol isn't defined in the loading module, then your app will be in a world of hurt, so care should be taken. -- Art Isbell NeXT/MIME Mail: arti@lavaDOTnet Trego Systems (for whom I don't speak) Voice/Fax: +1 808 394 0511 OPENSTEP/NT Voice Mail: +1 808 394 0495 managed care solutions US Mail: Honolulu, HI 96825-2638
From: drifterusa@macconnectNOT.com (John Bauer) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Thu, 11 Sep 1997 00:26:33 -0500 Organization: KAOS Message-ID: <199709110026331516569@accs-as12-dp07.dlls.grid.net> References: <5v1etj$84r$1@news.fas.harvard.edu> <LTaylor7-0909970929170001@vic-ca1-11.ix.netcom.com> Daniel L. Taylor wrote: > Steve's Fortune quote has me very doubtful and worried about his plans. He > stated that if he were in control he would milk the Mac for all it's worth > and move on to the next great thing. It's no secret that he and Ellison > believe NC's are the next great thing. One of the Mac sites had a link to an old MACWEEK article about Jobs's "fireside chat" at this year's WWDC in May. At that time, Jobs stated that if he were running Apple he would hand over hardware production to the cloners. Obviously he changed his mind. And this is a more recent quote than the FORTUNE article. Perhaps there's a lesson to be learned: Don't put too much faith in speculative thoughts from months or years past. > No, I shouldn't argue from my own personal hunch based on one Steve Jobs > quote. But often you don't have to look far to figure out what a person is > planning. Just read their lips.... Wasn't that FORTUNE article from early '96? That's looking pretty far if you ask me. -- John Bauer <remove NOT from email address above>
From: guyt@is.twi.tudelft.nl (A. Guyt) Newsgroups: comp.sys.next.programmer Subject: Rhapsody API's available here Date: 11 Sep 1997 08:01:00 GMT Organization: Delft University of Technology Message-ID: <5v88fs$erh$1@news.tudelft.nl> L.S., It is common knowledge that Apple recently WebPublished the new Rhapsody API's on their site. However as some of us may have experienced these docs cannot be unpacked by Openstep/Nextstep users. As the new Rhapsody API's have been significantly enhanced compared to the old Openstep API's I thought it might be useful to grab all the apple pieces together and bundle them in Openstep/Nextstep accessible format. # All information provided is owned & copyrighted by Apple, as soon as Apple objects to this service it will be interrupted # Foundation Kit classes + protocols + functions + types: http://dutifb.twi.tudelft.nl:8000/FoundationKit.tar.Z Application Kit classes + protocols + display postscript: http://dutifb.twi.tudelft.nl:8000/ApplicationKit.tar.Z May it lead to many good Rhapsody Apps !, Abraham Guyt. _____________________________________________________________________ Abraham Guyt P.O.Box 356 Department of Information Systems 2600 AJ Delft Faculty Information Technology & Systems The Netherlands Delft University of Technology tel: +31 15 278 5969 E-mail: guyt@is.twi.tudelft.nl NeXT-mail welcome
From: mpaque.spa-am@nospam.wco.com (Mike Paquette) Newsgroups: comp.sys.next.programmer Subject: Re: NSWindow hangs in new thread Date: Tue, 09 Sep 1997 19:53:14 GMT Organization: Electronics Service Unit No. 16 Message-ID: <341580e7.11380564@news.wco.com> References: <5uurvi$on5@dfw-ixnews12.ix.netcom.com> <34132b5f.3993642@news.wco.com> <5uvtp8$qbh@dfw-ixnews11.ix.netcom.com> <5v08pg$t4v@sjx-ixn7.ix.netcom.com> On 8 Sep 1997 07:17:04 GMT, mtrombin@ix.netcom.com (Mark Trombino) wrote: >So what have I learned? Only the thread that creates a window can remove it. > Now there's somethin' ya don't learn everyday! > Well, strictly speaking, thats true only on OPENSTEP for Windows, due to the use of Win32 windows. The limiting factor on Mach is the use of different DPS communications channels per thread. This makes each chanel thread safe, but if multiple threads are to manipulate the same graphics object, each thread must be careful to flush it's DPS context (really the buffered communications channel) before another thread does any DPS operations on the same object. (By the way, this is a problem with any graphics system which permits multiple pre-emptively scheduled threads to manipulate a graphics object. Some form of serialization must be done around all manipulations of the object in order to maintain a deterministic object state.) This gets sufficiently tricky and complex that for most simple multithreaded applications, just using Distributed Objects to serialize all drawing into a single thread suffices.
From: Arnaud Debayeux <debayeux@easynet.fr.NO_SPAM> Newsgroups: comp.sys.next.programmer Subject: Re: Custom message dispatching ? Date: Thu, 11 Sep 1997 14:49:27 +0200 Organization: [posted via] Easynet Group PLC Message-ID: <3417E8D7.BD0E65AE@easynet.fr.NO_SPAM> References: <3415D1F5.2A7102AA@easynet.fr.NO_SPAM> <5v71a6$120o$1@news.apple.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Mark Bessey wrote: > Wait, "externally to the object"? I don't get it. Do you mean that the > relationship is stored external to the object ? Yes. There is no code concerning the relationship in any object, everything is managed externally by the run-time. The object chosen as a source doesn't have any code to send a notification. It is the run-time which detects the event of a specific method invocation on a class or object, after that the run-time takes an action, like sending a message to a chosen destination object. The event, the source and dest class/object and the action are specified dynamically instead of being writen and compiled. > If you want a mechanism where, external to the implementation of either of > the > classes, you can say: "when this instance of the Mouse class receives the > set_x message, send a mouse_moved message to this instance of the Cursor > class", that's gonna be a challenge. > > > It's like the connected fields in Cosmo 3D/OpenGL Optimizer There is a Programmer's Guide Chapter 11 at http://www.sgi.com/Technology/openGL/optimizer/windows/index.html There are Input and Output Fields, and the standard actions are called Engines. This kind of 'channelization' is possible by at least 3DSMax and Softimage. How, I don't know. > I don't know that I like the idea of keeping relationship > information separate from the objects involved. It seems like you're moving > some information out of the class that it really needs to know about. How is it different to the 'relationship' writen explicitely in source code ? I want to avoid this static limitation of compilation, and allow a dynamic relationship management. I thought the OpenStep interactive demonstration (linking a scrollbar to a text-field to control a number value) was done this way. Well, it only creates code. This is what Meta-Object Protocol, Reflection, IntroSpection are about. > All the messaging in Objective-C goes through a > central point, a function called objc_msgSend(). You could modify that > function to support this functionality, I guess... I guess too, that's what I was asking for in my first message. I just obtained some source code of another run-time implementation (The Portable Object Compiler) to experiment with. The problem now is to have a good implementation like the method lookup cache table. Thanks for your comments.
From: dcoyle@ctp.com (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: Wait Cursor on NT Date: 11 Sep 1997 14:31:31 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5v8vc3$eb4$1@concorde.ctp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi all: I'm sure there's a FAQ for this, but I haven't found it. :-( I need to get a wait cursor going on NT, while I do fetches. I haven't bothered with - startWaitCursorTimer, since my gut feeling up front is that this won't work. ;-) dave -------------------------------------------------------------------------- /\/\ David A. Coyle, David A. Coyle, / /_ \ Cambridge Technology Partners, Cambridge Technology Partners, \ / / 118-119 Sráid Bhagoid Íocht, 118-119 Lower Baggot Street, \/\/ Baile Átha Cliath 2, Dublin 2, Phoblacht na hÉireann. Republic of Ireland. Guthán: +353 1 6079008 Tel: + 353 1 6079008 Greasán: dcoyle@ctp.com Fax: +353 1 6079001
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer Subject: Re: Rhapsody API's available here Date: 11 Sep 1997 14:10:18 GMT Organization: none Message-ID: <5v8u4a$8ie$1@ha2.rdc1.nj.home.com> References: <5v88fs$erh$1@news.tudelft.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: guyt@is.twi.tudelft.nl I downloaded these and put them on PEAK. They will, of course, be removed if Apple asks. You might want to try gzip rather than compress on them. Look at the difference % ls -l *.tar.Z -rw-rw-r-- 1 luomat user 458554 Sep 11 06:34 ApplicationKit.tar.Z -rw-rw-r-- 1 luomat user 420154 Sep 11 06:26 FoundationKit.tar.Z % gunzip -v *.Z ApplicationKit.tar.Z: -16.5% -- replaced with ApplicationKit.tar FoundationKit.tar.Z: 27.3% -- replaced with FoundationKit.tar (compress is so bad it actually made ApplicationKit.tar bigger!) % gzip -v *.tar ApplicationKit.tar: 20.7% -- replaced with ApplicationKit.tar.gz FoundationKit.tar: 49.8% -- replaced with FoundationKit.tar.gz % ls -l *.tar.gz -rw-rw-r-- 1 luomat user 311865 Sep 11 06:34 ApplicationKit.tar.gz -rw-rw-r-- 1 luomat user 290092 Sep 11 06:26 FoundationKit.tar.gz The gzipped versions can be found at ftp://next-ftp.peak.org/pub/openstep/new_arrivals/ or http://www.peak.org/openstep/new_arrivals/ where they are actually linked from /pub/openstep/multi-platforms/sourcelibrary/ TjL -- I'm sick of SPAM. Therefore, my email address is invalid. To email me, use one of these without spaces: luomat + next @ luomat.peak.org (for NeXT related stuff) luomat + peak @ luomat.peak.org (for PEAK related stuff)
From: simon@planon13.planon.qc.ca (Simon Glet) Newsgroups: comp.sys.next.programmer Subject: NSFormatter Date: 11 Sep 1997 14:45:45 GMT Organization: Internet-Login Message-ID: <5v906p$fjs$1@lamb.login.net> Hi, I'm having a big problem with NSFormatter. To format data to appear in a cell of a NSTableView, i subclassed NSFormatter applied it to the desired cell and it works fine for displaying the formatted data. When the user edits the cell and enters a correct value, it's almost impossible to stop editing the content of the cell ! I've searched the Web for some code examples using NSFormatter without success. I'm currently using OS 4.2 for MACH. Any pointers would be greatly appreciated. Thanks Simon Glet
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: Custom message dispatching ? Date: 11 Sep 1997 19:29:08 GMT Organization: Apple Computer, Inc. Message-ID: <5v9gq4$t1g$1@news.apple.com> References: <3417E8D7.BD0E65AE@easynet.fr.NO_SPAM> Arnaud Debayeux <debayeux@easynet.fr.NO_SPAM> writes > Mark Bessey wrote: > > > Wait, "externally to the object"? I don't get it. Do you mean that the > > relationship is stored external to the object ? > > Yes. There is no code concerning the relationship in any object, > everything is managed externally by the run-time. The object chosen as a > source doesn't have any code to send a notification. It is the run-time > which detects the event of a specific method invocation on a class or > object, after that the run-time takes an action, like sending a message > to a chosen destination object. The event, the source and dest > class/object and the action are specified dynamically instead of being > writen and compiled. Nifty. This still seems strange to me, though. And I can't see how it can be implemented easily in any language that I'm familiar with. Seems like you'd need a hook in the runtime that's checked *after* the method invocation on an object. Simple enough to do, I guess - a one-line change in objc_msgSend, and one more instance variable in Object, indicating whether this object has any "observers". > It's like the connected fields in Cosmo 3D/OpenGL Optimizer > > There is a Programmer's Guide Chapter 11 at > http://www.sgi.com/Technology/openGL/optimizer/windows/index.html > There are Input and Output Fields, and the standard actions are called > Engines. This kind of 'channelization' is possible by at least 3DSMax > and Softimage. How, I don't know. Unfortunately, I couldn't find the document you refer to, but this sounds a lot like the way outlets are handled in OPENSTEP (see below). I wonder if you're not accidentally making this a lot harder than it has to be... > > I don't know that I like the idea of keeping relationship > > information separate from the objects involved. It seems like you're > > moving some information out of the class that it really needs to know > > about. > > How is it different to the 'relationship' writen explicitely in source > code ? I want to avoid this static limitation of compilation, and allow > a dynamic relationship management. Well, if I'm trying to read a program (or heaven forbid, debug it), it's not going to be obvious where these extra triggered events are coming from if I haven't used this kind of system before. This is probably just my natural aversion to new things coming out again :-) > I thought the OpenStep interactive demonstration (linking a scrollbar to > a text-field to control a number value) was done this way. Well, it only > creates code. No, it doesn't create code. The way it works is like this: each control has an instance variable called "target" and one called "action". InterfaceBuilder just connects the "target" to whatever object you choose, and sets the "action" to a message to be sent to the "target" when the state of the control changes. However, each control can have only one target and one action. You can't specify an arbitrary number of targets, each with a different action. And each class has an instance variable that stores the target and action... > This is what Meta-Object Protocol, Reflection, IntroSpection are about. I don't think so. Introspection/Reflection usually refer to the ability to query objects at runtime about their capabilities, not the ability to insert arbitrary hooks into the message dispatch process. > > All the messaging in Objective-C goes through a > > central point, a function called objc_msgSend(). You could modify that > > function to support this functionality, I guess... > > I guess too, that's what I was asking for in my first message. > I just obtained some source code of another run-time implementation (The > Portable Object Compiler) to experiment with. The problem now is to have > a good implementation like the method lookup cache table. Good luck. Lets us know what you come up with, eh? (Although comp.lang.objective-c might be a more appropriate place for these discussions) -Mark -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: "Lawson English" <english@primenet.com> Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 10 Sep 1997 19:06:01 -0700 Organization: Primenet Services for the Internet Message-ID: <B03CA575-6110C@206.165.43.104> References: <5v6uq7$5na$2@nntp2.ba.best.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Delivery-Date: Wed, 10 Sep 1997 19:12:16 -0700 Return-Path: root@nntp02.primenet.com for <gherbert@crl.com>; Wed, 10 Sep 1997 19:05:51 -0700 (PDT) env-from (root@nntp02.primenet.com) by usr03.primenet.com (8.8.5/8.8.5) with ESMTP id TAA23193 for <sci-space-tech@moderators.uu.net>; Wed, 10 Sep 1997 19:10:14 -0700 (MST) for <sci-space-tech@moderators.uu.net>; Wed, 10 Sep 1997 19:10:15 -0700 (MST) id TAA01545 for sci-space-tech@moderators.uu.net; Wed, 10 Sep 1997 19:10:11 -0700 (MST) To: sci-space-tech@ncren.net nntp://news.primenet.com/comp.sys.mac.advocacy, nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.system, nntp://news.primenet.com/comp.sys.mac.misc Nathan Keir Edel <edel@best.com_SPAMBLOCK> said: > > The PowerPC chips are descended from the Power-II chipset used in older > IBM > RS/6000 systems. The PowerPC ISA is derived from the POWER ISA, not Power-II. I think that Power-II has more in common with the PowerPC instruction set than the POWER instruction set, although I haven't worked with it, so corrections are welcome. ------------------------------------------------------------------------- Build a better mousetrap, yes, but then market it like crazy, or instead of catching mice you'll only collect dust. -David Yeargin -------------------------------------------------------------------------
From: dental@precipice.com (Rick Sanford) Newsgroups: comp.sys.next.programmer Subject: Re: Automatic spinning cursor of Mach not going away ?? Date: 12 Sep 1997 02:03:32 GMT Organization: Dental Records (R) Message-ID: <5va7tk$er9@news1-alterdial.uu.net> References: <5uje9s$1q2$1@concorde.ctp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: tsengel@immd9.informatik.uni-erlangen.de In <5uje9s$1q2$1@concorde.ctp.com> Thomas Engel wrote: > Hi, > > NeXTs automatic wait cursor (for which they even own a patent!) is quite cool > and usually just does the right thing. > > But for one of my applications, under certain yet unknown circumstances, it > starts spinning...and never stops. > The application is fully active and works just fine..but the cursor is > spinning and I have not found an obvious way to reset it. > Any ideas ? > > Aloha > Tomi > don't know about that, but I have a case where in a certain app it occasionally stops spinning and stays on the screen as a colourful disk - until you nudge the mouse or hit a key, which makes it an arrow again. scary! makes you think it's locked-up. but noooo. I told the vendor but they don't know either. -- Rick Sanford Dental Records(R) dental@precipice.com NeXTMAIL welcome http://www.dentalrecords.com
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer Subject: Corrupted archive (was Rhapsody API's available here) Date: 12 Sep 1997 02:50:26 GMT Organization: none Message-ID: <5vaali$gjm$2@ha2.rdc1.nj.home.com> References: <5v9mkf$5m9$4@ha2.rdc1.nj.home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit There seems to be some problem with the archives: gnutar: archive ApplicationKit.tar.Z EOF not on block boundary So I've removed them from Peak. TjL -- I'm sick of SPAM. Therefore, my email address is invalid. To email me, use one of these without spaces: luomat + next @ luomat.peak.org (for NeXT related stuff) luomat + peak @ luomat.peak.org (for PEAK related stuff)
From: Oleg Zabluda <zabluda@math.psu.edu> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc,sci.space.policy Subject: Re: CPUs in Space (Was Re: No more PowerPCs... Date: 12 Sep 1997 03:16:08 GMT Organization: Penn State University, Center for Academic Computing Message-ID: <5vac5o$12cc@r02n01.cac.psu.edu> References: <5uujej$2to$1@nntp.ucs.ubc.ca> <B039C5D9-1E794@206.165.44.76> <5v2j5r$tqp$1@hecate.umd.edu> <5v563v$bsn@valhalla.comshare.com> Michael Pelletier <mikep@comshare.com> wrote: : ECC in the Intel platform detects two-bit errors and : corrects one, by using 12 (I think) Eight. : parity bits for 64 data bits. Oleg. -- Life is a sexually transmitted, 100% lethal disease.
From: Chris Johnson <cjohnson@object-works.com> Newsgroups: comp.sys.next.programmer Subject: Re: Wait Cursor on NT Date: Thu, 11 Sep 1997 23:59:44 -0400 Organization: Erol's Internet Services Message-ID: <3418BE30.6373EECE@object-works.com> References: <5v8vc3$eb4$1@concorde.ctp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit David, I have a category on NSCursor +(NSCursor *)waitCursor that creates an instance of an NSCursor using MS function calls to their wait cursor. In my app I then call methods in the App delegate -activateWaitCursor & -deactivateWaitCursor to push and pop this cursor off the cursor stack. This works great except if after you -activateWaitCursor a modal window session is invoked. If you must throw up an open, save or alert panel then you are going to have to -activateWaitCursor again as the cursor gets pushed off the stack. I am at home and works network is not connected to the internet so will try to follow this up with the source on the NSCursor category probably tomorrow. Chris Johnson David A. Coyle wrote: > Hi all: > > I'm sure there's a FAQ for this, but I haven't found it. :-( > > I need to get a wait cursor going on NT, while I do fetches. > I haven't bothered with - startWaitCursorTimer, since my gut feeling up front > is that this won't work. ;-) > > dave > > -------------------------------------------------------------------------- > /\/\ David A. Coyle, David A. Coyle, > / /_ \ Cambridge Technology Partners, Cambridge Technology Partners, > \ / / 118-119 Sráid Bhagoid Íocht, 118-119 Lower Baggot Street, > \/\/ Baile Átha Cliath 2, Dublin 2, > Phoblacht na hÉireann. Republic of Ireland. > > > Guthán: +353 1 6079008 Tel: + 353 1 6079008 > > Greasán: dcoyle@ctp.com Fax: +353 1 6079001
From: Rene Berber <_r.berber@computer.org_> Newsgroups: comp.sys.next.programmer Subject: Re: How to tell if an app is leaking memory/sucking swap (when should be idle)? Date: 12 Sep 1997 06:55:59 GMT Organization: None Whatsoever Message-ID: <5vap1v$i3a$1@hp.fciencias.unam.mx> References: <5v7uel$fmm$1@ha1.rdc1.nj.home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: luomat+next@luomat.peak.org Timothy J. Luoma wrote: > > I've been using a few apps which I haven't been using before. > > One of them seems to make the swapfile grow even when it isn't doing > anything. > > Is there some way I can test to find which one it is? > > The last time I asked this, I got a lot of responses suggesting ``quit > OmniWeb'' or something else like that. None of these apps is OmniWeb, or any > other OmniGroup application. > > I'd just like a way that I, a user without source code, can run an app > (perhaps over a series of days) to see if it is using more memory than it > should be. > > This may be a simple question, but I don't know where to look. Is ``ps'' a > good reference? If so, what fields are and what are not? > > Thanks > You can use top (is in the archive) which shows the processes ordered by cpu usage. In your case look for the VSIZE column which is the amount of virtual memory used by the process and it is directly related to the swaping space used (but the swap file is compressed.) Happy hunting.
Subject: Re: Custom message dispatching ? Newsgroups: comp.sys.next.programmer References: <3417E8D7.BD0E65AE@easynet.fr.NO_SPAM> <5v9gq4$t1g$1@news.apple.com> In-Reply-To: <5v9gq4$t1g$1@news.apple.com> From: marco@sente.ch (Marco Scheurer) Organization: Sen:te Message-ID: <3418f32c.0@epflnews.epfl.ch> Date: 12 Sep 97 07:45:48 GMT On 09/11/97, Mark Bessey wrote: > >Arnaud Debayeux <debayeux@easynet.fr.NO_SPAM> writes >> Mark Bessey wrote: > [...] >> > I don't know that I like the idea of keeping relationship >> > information separate from the objects involved. It seems >> > like you're moving some information out of the class that >> > it really needs to know about. >> >> How is it different to the 'relationship' writen explicitely >> in source code ? I want to avoid this static limitation of >> compilation, and allow a dynamic relationship management. > > Well, if I'm trying to read a program (or heaven forbid, debug it), > it's not going to be obvious where these extra triggered events > are coming from if I haven't used this kind of system before. > This is probably just my natural aversion to new things coming > out again :-) > [...] I'd like to point out a mechanism provided by EOF (Enterprise Objects Framework) which is close to what Arnaux describes. In a nutshell: when edited, objects send the 'change' message to themselves: [self change]; This has the effect to send a notification to registered observers. The EOAssociation class (and its subclasses), much like Arnaux's 'relationships', are observers, and when notified, act on other objects using only their public interface. EOAssociations can be set in InterfaceBuilder, in a slightly different way from the target/action way. This is how a text field or a browser can be updated without registering to any notification, without knowing anything about the changed object, and without the changed object knowing anything about its observers. This is done by sending the [self change] message instead of automatically by the runtime, as it could be done. I don't know if this would be a good idea, but IMHO, this EOF mechanism should be included as in OPENSTEP/Rhapsody, so that it can be used not only in expensive "database" custom applications. I believe that one misunderstanding between Mark and Arnaux comes from the term 'relationship', as in 'entity/relationship', or relationships like associations, classifications or composition structures between classes. With _this_ meaning, it would certainly be a bad idea to store relationships outside of objects: it breaks encapsulation, and it is a step back from the object model to the relational model. -- Marco Scheurer (marco@sente.ch) Sen:te
From: arti@lava.DOTnet (Art Isbell - remove "DOT") Newsgroups: comp.sys.next.programmer Subject: Re: Wait Cursor on NT Date: 12 Sep 1997 07:46:01 GMT Organization: LavaNet, Inc. Distribution: world Message-ID: <5varvp$fjr@mochi.lava.net> References: <5v8vc3$eb4$1@concorde.ctp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit dcoyle@ctp.com (David A. Coyle) wrote: > I need to get a wait cursor going on NT, while I do fetches. > I haven't bothered with - startWaitCursorTimer, since my gut feeling up front > is that this won't work. ;-) #import <System/Windows.h> #import <System/Winnt.h> SetCursor(LoadCursor(NULL, IDC_WAIT)) Doesn't this make you just want to get into more Win32 programming ?-) Looks like a lotta fun. "NT" stands for "New Technology", right? -- Art Isbell NeXT/MIME Mail: arti@lavaDOTnet Trego Systems (for whom I don't speak) Voice/Fax: +1 808 394 0511 OPENSTEP/NT Voice Mail: +1 808 394 0495 managed care solutions US Mail: Honolulu, HI 96825-2638
From: dcoyle@ctp.com (David A. Coyle) Newsgroups: comp.sys.next.programmer Subject: Re: Wait Cursor on NT Date: 12 Sep 1997 10:30:12 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5vb5jk$7j6$1@concorde.ctp.com> References: <5v8vc3$eb4$1@concorde.ctp.com> <5varvp$fjr@mochi.lava.net> arti@lava.DOTnet (Art Isbell - remove "DOT") wrote: >#import <System/Windows.h> >#import <System/Winnt.h> > > SetCursor(LoadCursor(NULL, IDC_WAIT)) > > Doesn't this make you just want to get into more Win32 programming ?-) >Looks like a lotta fun. "NT" stands for "New Technology", right? Yep. ;-) Now, how do I turn it off? :-) Dave (IDC_ARROW, I know...)
Date: Fri, 12 Sep 1997 07:02:06 -0600 From: rendto@icos.lol.li Subject: Communication problems from OS App via ORB/DOLE to Excel 97 Newsgroups: comp.sys.next.programmer Message-ID: <874065119.28574@dejanews.com> Organization: Deja News Posting Service Hello all, we 've used excel 95 for reporting output of our OS App with LocalID hex 09 (general english), everything looks right. Now after update to excel 97 many problems occur in communication with LocalID hex 09. As we switch to hex 409 (north american english) some function calls seems to work but all witch get a argument (like Range(..) ) does not do the right work! We get errors like 'method or attribute not recognized'. If we test these functions from within a visual basic program no problems occur. Perhaps we need a new Next ORB or nxorb.m file? Could anyone help? Torsten Rendelmann ICOS AG; FL-9490 Vaduz Tel. +41 75 23766-17 Fax +41 75 23766-34 mailto:rendto@icos.lol.li -------------------==== Posted via Deja News ====----------------------- http://www.dejanews.com/ Search, Read, Post to Usenet
Subject: Re: Custom message dispatching ? Newsgroups: comp.sys.next.programmer References: <3417E8D7.BD0E65AE@easynet.fr.NO_SPAM> <5v9gq4$t1g$1@news.apple.com> <3418f32c.0@epflnews.epfl.ch> In-Reply-To: <3418f32c.0@epflnews.epfl.ch> From: marco@sente.ch.mil (Marco Scheurer) Organization: Sen:te Message-ID: <3419309d.0@epflnews.epfl.ch> Date: 12 Sep 97 12:07:57 GMT On 09/12/97, Marco Scheurer wrote: >On 09/11/97, Mark Bessey wrote: > >I'd like to point out a mechanism provided by EOF (Enterprise Objects >Framework) which is close to what Arnaux describes. In a nutshell: >when edited, objects send the 'change' message to themselves: > > [self change]; > Sorry to follow up on my own post, but oops: this is [self willChange] and not [self change]. -- Marco Scheurer Sen:te (remove "dot mil" from my address for email)
From: Robert Forsyth <bobbyf@forsee.tcp.co.uk> Newsgroups: comp.sys.next.programmer,comp.lang.objective-c Subject: Re: Exception processing in Objective C? Date: Fri, 12 Sep 1997 15:50:34 +0100 Organization: Little Message-ID: <341956BA.72C9B63@forsee.tcp.co.uk> References: <340C6C07.5727@win.bright.net> <5v1f07$loh@shelob.afs.com> <873800789.32459@dejanews.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit David Stes wrote: > > Rich Schroedel <richs@win.bright.net> writes > > I'm in the process of rewriting some C++ code so that it can eventually > > be converted to Objective C. I'm stumped at how to handle exception > > processing. How do Objective C programmers handle the sort of situations > > that C++ programmers handle with exception processing? Any help out > > there? > > I think this can best be done by using Objective-C Blocks. > > For example, to handle "out of memory" in +new, (if you would not use > -gc), > > [{ > myObject = [Object new]; /* get an instance */ > } ifException: { > id exception | /* Block with one argument */ > [uselessObjects free]; /* free some memory */ > [exception resume]; /* "resume at location where failed */ > }]; > > This (specific) exception handler actually doesn't long-jump out of +new, > because of the |resume| message. If malloc() (inside +new) returns > NULL, then the handler will be evaluated. This particular handler frees > some memory, then +new resumes, and will attempt to malloc() again > (see below) until it succeeds. > > Of course, the handler could, instead of recovering, just fall > through (ie. long jump towards -ifException:) or reraise the exception, > which would evaluate the default handler, a Block that sends an -error: > message (aborts the process) saying "out of memory in +new" : > > + new > { > id newObject; > static id exBlock; > > /* exBlock needs to be instantiated before we're running out memory */ > if (!exBlock) exBlock = [{ [self error:"out of memory in +new"] } self]; > > while ((newObject = (*_alloc)(self,0)) == nil) [exBlock raise]; > return newObject; > } > > You can use Blocks for other things as well, like: > > id filtered; > id aCltn = [Collection new]; > ... > filtered = [aCltn collect:{id e | [set find:e]}]; > > which builds a Collection where it's rejecting the elements that > it does not find in "set", but anyhow, exception handling is > probably the most serious application. > > David. > > -------------------==== Posted via Deja News ====----------------------- > http://www.dejanews.com/ Search, Read, Post to Usenet Nextstep had serveral macros to enclose and handle exceptions a bit like Java, I can not remember the actual names of the macros! START_HANDLED_SECTION ... the bit that might RAISE an exception EXCEPTION_HANDLER ... the bit that handles the exception END_EXCEPTION_HANDLER Sorry if this is not much help, you will have to look in the manual to find the correct macro names! -- Rob. #:^) mailto://bobby@forsee.tcp.co.uk/ Please: NO SPAM NO JUNK E-MAIL or I will have to do something annoying.
From: Jim Redman <Jim@ergotech.com> Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Linking error on bundle Date: Fri, 12 Sep 1997 11:26:08 -0600 Organization: ErgoTech Systems, Inc. Message-ID: <34197B30.2EC7@ergotech.com> References: <34142B50.41C6@ina.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Philippe Converset <plc@ina.fr> Philippe Converset wrote: > An application with a single class "ClassA" an a loadable bundle with a > single class "ClassB". ClassB is using ClassA (sending a message in one > of its methods). > > The result of compilation is : > ClassB.o : error LNK2001 unresolved external symbol > .objc_class_name_ClassA > > It works fine under Mach but not under NT. > Did I forget a compiler or linker flag? > Is NT not supporting loadable bundles? -undefined suppress You will get a warning saying that the final target may not be valid. In the case you give this is untrue and the .dll will work correctly. _IF_ you reference functions or variables that are not linked to this bundle, for example functions or variables that are part of the executable file then these will not be resolved correctly and the bundle will crash when you try to access them. The workaround for this is to put these variables or functions in a framework and link this with the bundle. Jim
From: Jim Redman <Jim@ergotech.com> Newsgroups: comp.sys.next.programmer Subject: Compiling w/ SMP on NT Date: Fri, 12 Sep 1997 11:31:02 -0600 Organization: ErgoTech Systems, Inc. Message-ID: <34197C56.6BBE@ergotech.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Is there a way to improve performance of a compile using multiple processors under NT? My experience is that the compilation still uses only one processor. The -j option to make would seem to be one way to accomplish better utilization of all the CPUs, but doesn't seem to work under NT. Jim
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.newton.announce,comp.sys.newton.marketplace,comp.sys.newton.misc,comp.sys.newton.programmer,comp.sys.next.advocacy,comp.sys.next.announce,comp.sys.next.bugs,comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.northstar,comp.sys.nsc.32k,comp.sys.oric,comp.sys.palmtops,comp.sys.pen,comp.sys.powerpc,comp.sys.powerpc.advocacy,comp.sys.powerpc.misc,comp.sys.powerpc.tech,comp.sys.prime,comp.sys.proteon,comp.sys.psion,comp.sys.psion.announce,comp.sys.psion.apps,comp.sys.psion.comm,comp.sys.psion.marketplace,comp.sys.psion.misc,comp.sys.psion.programmer,comp.sys.psion.reviews,comp.sys.pyramid,comp.sys.ridge,comp.sys.sequent,comp.sys.sgi.admin,comp.sys.sgi.announce,comp.sys.sgi.apps,comp.sys.sgi.audio,comp.sys.sgi.bugs,comp.sys.sgi.graphics,comp.sys.sgi.hardware,comp.sys.sgi.marketplace,comp.sys.sgi.misc,comp.sys.sinclair,comp.sys.stratus,comp.sys.sun.admin Subject: cmsg cancel <34196E66.6C22@profit.com> Control: cancel <34196E66.6C22@profit.com> Date: 12 Sep 1997 16:33:20 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.34196E66.6C22@profit.com> Sender: Jimmy <jimmybob@profit.com> Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: nagle@netcom.com (John Nagle) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Fri, 12 Sep 1997 05:33:27 GMT Organization: Netcom On-Line Services Message-ID: <nagleEGDrFr.1Hw@netcom.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <5ur2uv$r3s@crcnis3.unl.edu> <5uujej$2to$1@nntp.ucs.ubc.ca> <5v2970$719@peabody.colorado.edu> <5v56db$c09@valhalla.comshare.com> Delivery-Date: Thu, 11 Sep 1997 22:30:53 -0700 Return-Path: nagle@netcom.com for <gherbert@crl.com>; Thu, 11 Sep 1997 22:28:45 -0700 (PDT) env-from (nagle@netcom.com) id QQdgqg25933; Fri, 12 Sep 1997 01:33:30 -0400 (EDT) mikep@comshare.com (Michael Pelletier) writes: >In article <5v2970$719@peabody.colorado.edu>, > Frank Crary <fcrary@rintintin.Colorado.EDU> wrote: >> >>Two comments. First, I though it was a RAD4000, which would be a >>radiation hardened version of the R4000 chip. Second, last time >>I checked, the PowerPC was a computer (or series of computers), >>not a chip. So the connection to PowerPCs is simply that they >>use the a similar processor. (Nor are PowerPC the only computers >>that use this family of chips: SGI uses them extensively; the >>Indigo II on my desk at work uses a R4400 processor.) I think it's >>important to avoid confusing chips with computers. >Nope, Frank -- the PowerPC is the name of the chip. Pop open your >Macintosh, or an IBM C30, or whatever, and take a look at the >CPU. It'll have a nifty little "PowerPC" logo on it. IBM has a line of computers called the "Power Series", which started as RISC workstations. The PowerPC is more or less compatible with that architecture. The MIPS machines, such as the R4000, are a completely different family of chips. MIPS is now owned by Silicon Graphics, which had to buy them to keep their CPU supplier from going away. John Nagle
From: bmeadows@acl.gatech.edu (Brian K. Meadows) Newsgroups: comp.sys.next.programmer Subject: Re: How to tell if an app is leaking memory/sucking swap (when should be idle)? Date: Fri, 12 Sep 1997 18:45:23 GMT Organization: Georgia Institute of Technology Message-ID: <34198d21.1373492@news.gatech.edu> References: <5v7uel$fmm$1@ha1.rdc1.nj.home.com> Open a shell window and enter the command: ps -aux|more . You'll get a list of CPU and VM loads by job number. It isn't as good as top, but it is a native command. Brian
From: "Daniel T. Fahey" <danfahey@dansources.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.marketplace Subject: Jobs: Next Developers Date: Fri, 12 Sep 1997 15:50:57 -0700 Organization: DanSources Technical Services Message-ID: <3419C751.1511@dansources.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit HELLO: We are seeking Junior Next Developer and a few intermediate and Senior Developers as well. Two positions are open for Team Leads, must have great Systems and Application Development skills and able to lead a team of up to five personality challenged hot shot Developers ;-) 2 Senior positions are in Philly and the rest of the mix in Northern Virginia. Give me a call Sincerely Dan Fahey -- For more information about DanSources Technical Services Incorporated Feel free to check out our Web Page at http://www.dansources.com // See Jobs
From: arti@lava.DOTnet (Art Isbell - remove "DOT") Newsgroups: comp.sys.next.programmer Subject: M-V-C design question Date: 12 Sep 1997 22:09:26 GMT Organization: LavaNet, Inc. Distribution: world Message-ID: <5vceim$evp@mochi.lava.net> I have a design question about appropriate Model-View-Controller Controller to Model communication. The rewrite caused by moving from DBKit to EOF gave me an opportunity to redesign an app that had not been designed originally with M-V-C separation in mind. The redesign feels much better than the old design because Model classes don't know anything about View or Controller classes and thus are much more reusable. View classes know nothing about underlying Model classes. Controller classes tie Views to Models. The advertised advantages of M-V-C seem apparent. Life is good (at least with the design :-) I have Model classes that implement a wide variety of functionality. Rather than writing one mongo Controller class that attempts to associate Views with the many Model classes, I wrote several Controller classes that are teamed with various Model classes based on functionality (e.g., sign on, security, licensing, window management, preferences, etc.). Does good M-V-C design dictate which Controllers should communicate with which Models, are is any communication pattern OK? Should Controllers in general communicate only with Models associated with them and with other Controllers, or should any Controller be able to communicate with any Model or Controller? Allow me to illustrate. Suppose Controller0 needs the services of Model1 which is "associated" with Controller1 by functionality. Should Controller0 communicate directly with Model1 or communicate via Controller1 to get Model1's services? Controller0 Controller1 \ \ \ \ Model0 Model1 or Controller0--->Controller1 | | | | Model0 Model1 The second approach results in more messages, but eliminates the dependency of Controller0 on Model1 because changes to Model1 don't affect Controller0. Controller1 is already dependent on Model1 because it created Model1 and is more closely related to it than is Controller0. So the second approach feels better. Changes in Model1 don't require Controller0 to be recompiled and just seems to isolate functionality better. -- Art Isbell NeXT/MIME Mail: arti@lavaDOTnet Trego Systems (for whom I don't speak) Voice/Fax: +1 808 394 0511 OPENSTEP/NT Voice Mail: +1 808 394 0495 managed care solutions US Mail: Honolulu, HI 96825-2638
From: andydunn@op.net (Andy Dunn) Newsgroups: comp.sys.next.programmer Subject: Re: Rhapsody API's available here Date: 12 Sep 1997 22:20:08 GMT Organization: OpNet -- Greater Philadelphia Internet Service Message-ID: <5vcf6o$a92$1@picasso.op.net> References: <5v88fs$erh$1@news.tudelft.nl> <5v8u4a$8ie$1@ha2.rdc1.nj.home.com> Cc: nospam@all.please In <5v8u4a$8ie$1@ha2.rdc1.nj.home.com> Timothy J. Luoma wrote: > > I downloaded these and put them on PEAK. They will, of course, be removed if > Apple asks. > ><<snipped>> I couldn't find them... Did Apple object that quickly, or are they not really posted yet?
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.next.programmer,comp.sys.mac.programmer.misc,comp.sys.next.misc,comp.sys.mac.hypercard Subject: What do Mac/NeXT/Rhapsody developers think about all this? Date: 12 Sep 1997 16:47:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B03F27C6-1A3543@206.165.44.45> nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.next.programmer, nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.misc, nntp://news.primenet.com/comp.sys.mac.hypercard MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit It's a good question, isn't it? To find out what Macintosh and NeXT hardware and software developers think (they may not know WHAT to think, BTW) about the issues that have popped up in the last week or two concerning clone-licensing, the new Apple-less IBM-Motorola initiatives concerning PowerPCs, etc., you should participate in/lurk-in the AIMED-talk mailing list. AIMED stands for Association of Independent Macintosh Engineers and Developers Subscription info is to be found below: Spread the word and feel free to participate yourself. Think of this as a town hall for Mac users and developers. ++++++++++++++++++ <http://www.aimed.org> ++++++++++++++ HOW DO I SUBSCRIBE TO AIMED-TALK? ------------------------------------- To subscribe to AIMED-TALK proceed as follows: Send E-mail to: <aimed-talk-request@aimed.org> with a blank subject line and the following command as the first (and only) line of the message body: SUBSCRIBE ++++++++++++++++++ --------------------------------------------------- As of 11 Sept 97, A.I.M. is now spelled "I.M." as in "I'm gone." ---------------------------------------------------
From: Chris Van Buskirk <cvbuskirk@home.com> Newsgroups: comp.sys.next.programmer Subject: rhapsody java apis... Date: Fri, 12 Sep 1997 23:40:58 -0400 Organization: @Home Networks Message-ID: <341A0B4A.FEF3347E@home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Anyone seen the java apis for rhapsody yet? A review would be cool. -chris
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer Subject: Re: Rhapsody API's available here Date: 13 Sep 1997 04:28:15 GMT Organization: none Message-ID: <5vd4ov$8b2$2@ha2.rdc1.nj.home.com> References: <5v88fs$erh$1@news.tudelft.nl> <5v8u4a$8ie$1@ha2.rdc1.nj.home.com> <5vcf6o$a92$1@picasso.op.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: andydunn@op.net In <5vcf6o$a92$1@picasso.op.net> Andy Dunn wrote: > > In <5v8u4a$8ie$1@ha2.rdc1.nj.home.com> Timothy J. Luoma wrote: > > > > I downloaded these and put them on PEAK. They will, of course, be removed > if > > Apple asks. > > > ><<snipped>> > > I couldn't find them... > > Did Apple object that quickly, or are they not really posted yet? They were corrupted, as I posted later on. I'm still waiting for a clean version (the versions on the original site were also corrupted, not just the peak version). TjL -- I'm sick of SPAM. Therefore, my email address is invalid. To email me, use one of these without spaces: luomat + next @ luomat.peak.org (for NeXT related stuff) luomat + peak @ luomat.peak.org (for PEAK related stuff)
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer Subject: Re: rhapsody java apis... Date: 13 Sep 1997 04:30:55 GMT Organization: none Message-ID: <5vd4tv$8b2$3@ha2.rdc1.nj.home.com> References: <341A0B4A.FEF3347E@home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: cvbuskirk@home.com In <341A0B4A.FEF3347E@home.com> Chris Van Buskirk wrote: > Anyone seen the java apis for rhapsody yet? A review would be cool. Even if they had, they'd be under NDAs I think.... > Organization: @Home Networks Seems like we're getting closer to a NUG of @Home customers! TjL, also using @Home -- I'm sick of SPAM. Therefore, my email address is invalid. To email me, use one of these without spaces: luomat + next @ luomat.peak.org (for NeXT related stuff) luomat + peak @ luomat.peak.org (for PEAK related stuff)
From: "Gregory H. Anderson" <greg@afs.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: Sat, 13 Sep 1997 01:41:10 -0400 Organization: Anderson Financial Systems Inc. Message-ID: <341A2776.150A@afs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I can't take credit for discovering this, but I'll be more than happy to take credit for popularizing it until the person who told me steps forward to share the glory... In Openstep 4.2 (which includes the Mach and Windows releases that were distributed as "Prelude to Rhapsody"), there is a header file in the AppKit framework called NSInterfaceStyle. If you're running on Windows, try this in an MSDOS window (proper capitalization is important): defaults write NSGlobalDomain NSInterfaceStyle NextStep then logout, log back in, and launch some Openstep apps. Surprise! If you're really into pain on a Mach-equipped box, the complementary command in a Terminal window is: defaults write NSGlobalDomain NSInterfaceStyle Windows95 As with most Openstep defaults, you can substitute an application name for "NSGlobalDomain", in which case only that app will change. If you get sick of this trick, type this to return to normal: defaults remove NSGlobalDomain NSInterfaceStyle Sure looks like the beginnings of an Appearance Manager to me. But what do I know, I'm just a simple country programmer... 8^) Greg PS: The reason I cross-posted this to the programmer groups was to make the point that you can test multiple interface styles and layouts and sizes, even you if you don't own all the supported equipment. Name another cross-platform environment where you can do _that_.
From: mikep@comshare.com (Michael Pelletier) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc,sci.space.policy Subject: Re: CPUs in Space (Was Re: No more PowerPCs... Date: 10 Sep 97 04:02:07 GMT Organization: Comshare, Inc. Message-ID: <5v563v$bsn@valhalla.comshare.com> References: <5uujej$2to$1@nntp.ucs.ubc.ca> <B039C5D9-1E794@206.165.44.76> <5v2j5r$tqp$1@hecate.umd.edu> Delivery-Date: Thu, 11 Sep 1997 04:08:06 -0700 Return-Path: uucp@marie.iijnet.or.jp for <gherbert@crl.com>; Thu, 11 Sep 1997 04:07:18 -0700 (PDT) env-from (uucp@marie.iijnet.or.jp) id QQdgnk13058; Thu, 11 Sep 1997 07:11:19 -0400 (EDT) To: sci-space-tech@uunet.uu.net In article <5v2j5r$tqp$1@hecate.umd.edu>, David T. Wang <davewang%wam.umd.edu.delete.after.edu@Glue.umd.edu> wrote: > >You are ofcourse correct. From what I understand, if you use one of the >older process technologies, like 2.0um, it has thick enough oxides so that >you won't easily get your oxides annihilated by cosmic ray. However, there's >not much you can do about single event upsets. (Cosmic ray goes through, and >charges/discharges a single bit, and flips it.) Of course there's something you can do about it -- use ECC RAM! By adding enough parity bits, you can detect and correct as many bit errors as you care to. ECC in the Intel platform detects two-bit errors and corrects one, by using 12 (I think) parity bits for 64 data bits. I knew the formula for determining parity needs for n-bit error correction, from a class I took, but I don't have the foggiest idea what it is now. It's been quite a while. -Mike Pelletier.
From: mikep@comshare.com (Michael Pelletier) Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 10 Sep 97 04:07:07 GMT Organization: Comshare, Inc. Message-ID: <5v56db$c09@valhalla.comshare.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <5ur2uv$r3s@crcnis3.unl.edu> <5uujej$2to$1@nntp.ucs.ubc.ca> <5v2970$719@peabody.colorado.edu> Delivery-Date: Thu, 11 Sep 1997 04:08:05 -0700 Return-Path: uucp@marie.iijnet.or.jp for <gherbert@crl.com>; Thu, 11 Sep 1997 04:07:16 -0700 (PDT) env-from (uucp@marie.iijnet.or.jp) id QQdgnk13062; Thu, 11 Sep 1997 07:11:20 -0400 (EDT) To: sci-space-tech@uunet.uu.net In article <5v2970$719@peabody.colorado.edu>, Frank Crary <fcrary@rintintin.Colorado.EDU> wrote: > >Two comments. First, I though it was a RAD4000, which would be a >radiation hardened version of the R4000 chip. Second, last time >I checked, the PowerPC was a computer (or series of computers), >not a chip. So the connection to PowerPCs is simply that they >use the a similar processor. (Nor are PowerPC the only computers >that use this family of chips: SGI uses them extensively; the >Indigo II on my desk at work uses a R4400 processor.) I think it's >important to avoid confusing chips with computers. Nope, Frank -- the PowerPC is the name of the chip. Pop open your Macintosh, or an IBM C30, or whatever, and take a look at the CPU. It'll have a nifty little "PowerPC" logo on it. See: <http://www.mot.com/SPS/PowerPC/products/semiconductor/chips.html> -Mike Pelletier.
From: mmalcolm crawford <malcolm@plsys.co.uk> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 13 Sep 1997 11:06:30 GMT Organization: P&L Systems Message-ID: <5vds3m$di4$1@ironhorse.plsys.co.uk> References: <341A2776.150A@afs.com> In-Reply-To: <341A2776.150A@afs.com> On 09/13/97, "Gregory H. Anderson" wrote: >I can't take credit for discovering this, but I'll be more than happy to >take credit for popularizing it until the person who told me steps >forward to share the glory... > Thanks for the lead-in Greg... :-) ... however (I'm really sorry about this) actually the first person to publicly announce this was Thomas Engel, a month ago: --- Subject: Rhapsody and UI Themes...they already work. From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Date: 1997/08/14 Newsgroups: comp.sys.next.advocacy,comp.sys.mac.advocacy while all this raving is going on abotu the leak Ui document...propably some folks should just realize that Openstep already supports two themes: - NextStep - Windows95 If you are running OPENSTEP/4.x Mach go to the Termnial and do: defaults write NSGlobalDomain NSInterfaceStyle Windows95 --- Doubtless, though, more on this subject anon! ;-) Best wishes, mmalc.
From: nurban@sps1.phys.vt.edu (Nathan Urban) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 13 Sep 1997 10:39:17 -0400 Organization: Data Systems Consulting, Inc. Message-ID: <5ve8il$g86$1@sps1.phys.vt.edu> References: <341A2776.150A@afs.com> In article <341A2776.150A@afs.com>, greg@afs.com wrote: > In Openstep 4.2 (which includes the Mach and Windows releases that were > distributed as "Prelude to Rhapsody"), there is a header file in the > AppKit framework called NSInterfaceStyle. If you're running on Windows, > try this in an MSDOS window (proper capitalization is important): > defaults write NSGlobalDomain NSInterfaceStyle NextStep What would be _really_ nice is if this worked in Rhapsody, so you could get the old look-and-feel instead of the MacOS Platinum..
From: dlbeatty@[nospam].texas.net Newsgroups: comp.sys.next.programmer Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 13 Sep 1997 15:10:47 GMT Organization: Texas Networking, Inc. Message-ID: <5veadn$so4$2@news3.texas.net> References: <341A2776.150A@afs.com> In Openstep 4.2 / Prelude to Rhapsody "Easter Egg" comp.sys.next.programmer "Gregory H. Anderson" <greg@afs.com> writes, > PS: The reason I cross-posted this to the programmer groups was > to make the point that you can test multiple interface styles > and layouts and sizes, even you if you don't own all the supported > equipment. Name another cross-platform environment where you can > do _that_. Visualworks\Smalltalk from ParcPlace used to support something like this, about 4 years ago.
From: alviani@ix.netcom.com (Frank Alviani) Newsgroups: comp.sys.next.programmer Subject: [Q] InterfaceBuilder Problem in P2R Date: Sat, 13 Sep 1997 11:16:27 -0500 Organization: Netcom Message-ID: <alviani-ya02408000R1309971116270001@nntp.netcruiser> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Hello all, I am working on my first non-tutorial project in OS 4.2 (Prelude to Rhapsody distribution) and have run into a problem. In the main nib file, I have instantiated 2 NSMutableArray objects, connected to the application controller instance. This was working for a while, but I just started getting the following error when building: [build window] Multiply defined symbols (see output text); link errors [WM console] InteraceBuilder [437] - () appears twice in object table I had re-imported the header file for the controller class after adding an outlet and several messages. I have been making changes in the header file and then updating the nib file as I evolve the object structure. My impression is that this is a fairly conventional approach to program development. My specific questions are: (1) How would I look at the "output text" (I believe this refers to the object file) to determine what symbol is multiply defined? I tried otool, but this is rather confusing. Double clicking on the error message does nothing. (2) If (as I suspect) the nib file is at fault, is it illegal to have multiple instantiations of 1 class in the same nib file. Each instance is connected to a distinct outlet in the controller class? (3) While directly instantiating classes in IB for non-interface objects looks like a nice technique (these are internal indices), is this an "accepted technique"? Should I instead be programatically creating them? The interface suggests that this should work. All help that you can provide will be greatly appreciated, as I cannot continue development until I get this problem resolved. Any pointers to clearer documentation will also be appreciated; what is provided online is spotty and assumes a great deal of existing UNIX knowledge (I've been a Mac programmer for the last 13 years - lots of OO experience, but no Unix). Thanks in advance, Frank Alviani
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 13 Sep 1997 16:47:33 GMT Organization: none Message-ID: <5veg35$12e$3@ha2.rdc1.nj.home.com> References: <341A2776.150A@afs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: greg@afs.com In <341A2776.150A@afs.com> "Gregory H. Anderson" wrote: > Sure looks like the beginnings of an Appearance Manager to me. But what > do I know, I'm just a simple country programmer... 8^) Since I'm stuck with lowly 4.1 rather than 4.2, I'm not sure what this means. It would be nice if it means that NeXTers will be able to make Rhapsody look like NS/OS 4.x if they want, and other OS users can customize to their lesser OSes ;-) TjL -- I'm sick of SPAM. Therefore, my email address is invalid.
Message-ID: <341AF521.6025DA8@neteffects.com> Date: Sat, 13 Sep 1997 13:18:41 -0700 From: "Jack G. Bader" <jbader@neteffects.com> Organization: NetEffects Inc. MIME-Version: 1.0 Subject: *** NeXTStep Wannabees - Entry level positions Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Newsgroups: comp.sys.next.programmer NetEffects Inc. has an immediate opening for entry-level Objective-C consultants to work with our telecommunications client in Minneapolis. You will join a team that is responsible for Business Analysis, Delvelopment and support of customer service applications. Your responsibilities will involve: - Maintaining Objective-C applications - Enhancing existing applications The technical environment includes NeXTStep, Objective-C, Smalltalk, WindowsNT and Oracle. If you have a degree in CS and any educational / professional experience with Objective-C, this would be an excellent opportunity for you to get into a professional organization. Please respond immediately with your resume. Our process is quite short and you will recieve quick feedback. Please respond to: Jack Bader jbader@neteffects.com 314-727-1107 314-444-6866 fax
Date: Sat, 13 Sep 1997 14:45:35 -0400 From: joe.ragosta@dol.net (Joe Ragosta) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Message-ID: <joe.ragosta-ya02408000R1309971445350001@news.dol.net> References: <341A2776.150A@afs.com> <5vds3m$di4$1@ironhorse.plsys.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit In article <5vds3m$di4$1@ironhorse.plsys.co.uk>, mmalcolm crawford <malcolm@plsys.co.uk> wrote: > while all this raving is going on abotu the leak Ui document...propably some > folks should just realize that Openstep already supports two themes: > > - NextStep > - Windows95 > > If you are running OPENSTEP/4.x Mach go to the Termnial and do: > > defaults write NSGlobalDomain NSInterfaceStyle Windows95 If you can think of any good reason to _want_ the Win95 interface, that is. -- Regards, Joe Ragosta joe.ragosta@dol.net Visit the Complete Macintosh Web Site http://www.dol.net/~Ragosta/complmac.htm
From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 13 Sep 1997 20:06:56 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5verp0$f7i$1@concorde.ctp.com> References: <341A2776.150A@afs.com> <5vds3m$di4$1@ironhorse.plsys.co.uk> <joe.ragosta-ya02408000R1309971445350001@news.dol.net> joe.ragosta@dol.net (Joe Ragosta) wrote: >In article <5vds3m$di4$1@ironhorse.plsys.co.uk>, mmalcolm crawford ><malcolm@plsys.co.uk> wrote: > >> while all this raving is going on abotu the leak Ui document...propably some >> folks should just realize that Openstep already supports two themes: >> >> - NextStep >> - Windows95 >> >> If you are running OPENSTEP/4.x Mach go to the Termnial and do: >> >> defaults write NSGlobalDomain NSInterfaceStyle Windows95 > >If you can think of any good reason to _want_ the Win95 interface, that is. > I have one...a really good one :-) Our software _must_ run on WinNT once it ships...but _none_ of the core developers was willing to develop on NT... ...so in order to test final look (some elemets change in size in the Win theme...colors need to match etc) we simple run these apps with Win Interface. A really good reason if you ask me...you don't have to touch NT at all. Only the guys we do system test have to run that thing. Aloha Tomi P.S. Want to hear some NT stories ? about swap files etc.. should we open a new newsgroup for that. comp.sys.next.winstories :-)
From: bettis@inetnebr.com (Mr. Jeremy Bettis) Newsgroups: comp.sys.next.programmer Subject: Re: Wait Cursor on NT Date: 13 Sep 1997 15:56:21 -0500 Organization: Internet Nebraska Message-ID: <5veull$1fs$1@falcon.inetnebr.com> References: <5v8vc3$eb4$1@concorde.ctp.com> <5varvp$fjr@mochi.lava.net> <5vb5jk$7j6$1@concorde.ctp.com> NNTP-Posting-User: bettis dcoyle@ctp.com (David A. Coyle) writes: >arti@lava.DOTnet (Art Isbell - remove "DOT") wrote: >>#import <System/Windows.h> >>#import <System/Winnt.h> >> >> SetCursor(LoadCursor(NULL, IDC_WAIT)) >> >> Doesn't this make you just want to get into more Win32 programming ?-) >>Looks like a lotta fun. "NT" stands for "New Technology", right? >Yep. ;-) >Now, how do I turn it off? :-) >Dave >(IDC_ARROW, I know...) It will turn itself off when the next window event is processed. The only time you need to turn it off by hand is when it turned on as a result of a menu click. That has been my experience.
From: jonwhite@erols.com (J. White) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sat, 13 Sep 1997 17:56:09 -0400 Organization: Erol's Internet Services Message-ID: <jonwhite-1309971756090001@phd-as2s21.erols.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <19970907234113405260@accs-as15-dp07.dlls.grid.net> In article <19970907234113405260@accs-as15-dp07.dlls.grid.net>, drifterusa@macconnectNOT.com (John Bauer) wrote: > > This is the point: *very few* companies chose Macintoshes despite the > platform's superior ease of use and inherent long-term cost advantages. > Apple's mounting credibility problems may make even fewer consider > Rhapsody no matter how wonderful it is. As the history of Apple shows > time and again, perception outweighs technological merits. > Your absolutely right. Perception and marketing sells products. Apple has always been pathetic with their marketing. And now the perception factor is against them in IS departments. My company uses mac office machines and unix boxes to run our business. With Apple's recent moves, now I must ask myself if I'm comfortable buying Mac systems, possibly jeopardizing our companies competitiveness. I've decided that this monday, I'll call Apple and ask them why I should keep buying their machines. There answer had better be very good. Jon White > -- > John Bauer <remove NOT from email address above>
From: nurban@sps1.phys.vt.edu (Nathan Urban) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 13 Sep 1997 19:02:42 -0400 Organization: Data Systems Consulting, Inc. Message-ID: <5vf62i$iju$1@sps1.phys.vt.edu> References: <341A2776.150A@afs.com> <5veg35$12e$3@ha2.rdc1.nj.home.com> In article <5veg35$12e$3@ha2.rdc1.nj.home.com>, nospam@all.please (Timothy J. Luoma) wrote: > In <341A2776.150A@afs.com> "Gregory H. Anderson" wrote: > > Sure looks like the beginnings of an Appearance Manager to me. But what > > do I know, I'm just a simple country programmer... 8^) > Since I'm stuck with lowly 4.1 rather than 4.2, I'm not sure what this means. The NSInterfaceStyle default also works under 4.1. Try it!
From: Sherwood Glazier <sglazier@wlrk.com> Newsgroups: comp.sys.next.programmer Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: Sat, 13 Sep 1997 18:54:52 -0400 Organization: Wachtell, Lipton, Rosen & Katz Message-ID: <341B19BC.1FE1@wlrk.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit In article <5vds3m$di4$1@ironhorse.plsys.co.uk>, mmalcolm crawford <malcolm@plsys.co.uk> wrote: > while all this raving is going on abotu the leak Ui document...propably some > folks should just realize that Openstep already supports two themes: > > - NextStep > - Windows95 > > If you are running OPENSTEP/4.x Mach go to the Termnial and do: > > defaults write NSGlobalDomain NSInterfaceStyle Windows95 How do you change it back to the NeXT interface? --- defaults write NSGlobalDomain NSInterfaceStyle NeXT ??? Just curious. Sherwood Glazier sglazier@wlrk.com
From: Bob Korsan <bobk@decide.com> Newsgroups: comp.sys.next.programmer Subject: Replacing hard disk on Intel NS 3.2 Date: Sat, 13 Sep 1997 16:46:06 +0000 Organization: Decisions, Decisions! Message-ID: <341AC34E.159E@decide.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit CC: bobk@decide.com Hi, I have had a NS 3.2 system for several years now (I'm not a developer, just a Mathematica User and NS seemed to be the best OS). The disk drive is now making wierd noises and I want to replace it before it crashes. I have a 2 MB DAT drive which does a backup every night, so the info should be there... If I go out and buy a new drive and install it, what is the procedure to get 3.2 up and running again? The hardware vendor did the initial install, so I have never done anything except run the box. (Oh yes, this was an educational version of the OS so I have only the most minimal of documentation -- I was teaching but have since stopped. I don't have help from the campus computer support like I did before.) Can anyone lay out the steps that I would have to follow? Can anyone point me to a web site or ftp site which would have the info I would like? (I tried NeXTanswers, but after about 2-3 hours of wandering and searching that site, I couldn't find anything. I sent a message to Apple Enterprise Support, but they say the OS is no longer supported. They recommended I message this group, so I am.) Lastly, I haven't had any trouble with the OS, so I don't see the need to upgrade, but if there is something really wiz bang on the latest version, could someone let me know what I would be getting? And if you know the price to upgrade from 3.2 to the latest (I don't even know what the latest version is), please let me know. Thanks in advance for your help. Peace/Bob
From: arti@lava.DOTnet (Art Isbell - remove "DOT") Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 13 Sep 1997 23:52:36 GMT Organization: LavaNet, Inc. Distribution: world Message-ID: <5vf904$ahr@mochi.lava.net> References: <341A2776.150A@afs.com> "Gregory H. Anderson" <greg@afs.com> wrote: > In Openstep 4.2 (which includes the Mach and Windows releases that were > distributed as "Prelude to Rhapsody"), there is a header file in the > AppKit framework called NSInterfaceStyle. If you're running on Windows, > try this in an MSDOS window (proper capitalization is important): > > defaults write NSGlobalDomain NSInterfaceStyle NextStep > > then logout, log back in, and launch some Openstep apps. Surprise! > PS: The reason I cross-posted this to the programmer groups was to make > the point that you can test multiple interface styles and layouts and > sizes, even you if you don't own all the supported equipment. Name > another cross-platform environment where you can do _that_. Doesn't really work very will under OS/NT because the windows are true Windows GDI, not DPS, windows, so the Windows 95 menus remain across the top of each window instead of using a separate NEXTSTEP menu. This increases the height of each window. Also, when specifying the NEXTSTEP interface style and opening one of our windows that contains a tableView subclass that works fine under NT, the app crashed with a PostScript error, so more work remains. -- Art Isbell NeXT/MIME Mail: arti@lavaDOTnet Trego Systems (for whom I don't speak) Voice/Fax: +1 808 394 0511 OPENSTEP/NT Voice Mail: +1 808 394 0495 managed care solutions US Mail: Honolulu, HI 96825-2638
From: nebula128@aol.com (Nebula128) Newsgroups: comp.sys.next.programmer Subject: BASIC HELP... PLEASE!!!!!!! Date: 13 Sep 1997 23:58:25 GMT Message-ID: <19970913235800.TAA12816@ladder01.news.aol.com> Organization: AOL http://www.aol.com SnewsLanguage: English I'm beginning to program in BASIC at my high school. I was wondering if there was any basic compiler that creats .exe files. Will Visual BASIC do that?? Or does EVERY program i write have to be run in a QBASIC or similar enviroment. By the way, What can VB do??? Thanks for answering! PS: I'm not a novice or anything, i know Pascal too. You'll be helping a equal. :) ~~~~~Beyond Imagination Lies The Truth...~~~~~ ~~~~~~~~~~Nebby@pilotmail.net~~~~~~~~~~~ :)
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 14 Sep 1997 00:04:54 GMT Organization: Anderson Financial Systems Inc. Distribution: world Message-ID: <5vf9n6$j4m@shelob.afs.com> References: <5vf904$ahr@mochi.lava.net> Art Isbell - remove "DOT" writes > Doesn't really work very will under OS/NT because the windows are > true Windows GDI, not DPS, windows, so the Windows 95 menus remain > across the top of each window instead of using a separate NEXTSTEP menu. > This increases the height of each window. > > Also, when specifying the NEXTSTEP interface style and opening one > of our windows that contains a tableView subclass that works fine under > NT, the app crashed with a PostScript error, so more work remains. Like I said, it's a good _beginning._ -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 14 Sep 1997 00:05:47 GMT Organization: Anderson Financial Systems Inc. Message-ID: <5vf9or$j4p@shelob.afs.com> References: <341B19BC.1FE1@wlrk.com> Sherwood Glazier <sglazier@wlrk.com> writes > How do you change it back to the NeXT interface? > > --- defaults write NSGlobalDomain NSInterfaceStyle NeXT ??? Use "NextStep", or do a "defaults remove". If there is no default in the database, it uses the native format. -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
From: Ronald Pomeroy <rpomeroy@dallas.net> Newsgroups: comp.sys.next.programmer Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: Sat, 13 Sep 1997 19:01:51 -0500 Organization: D'A Message-ID: <341B296E.F2F4AEF8@dallas.net> References: <341A2776.150A@afs.com> <5veadn$so4$2@news3.texas.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit dlbeatty@[nospam].texas.net wrote: > In Openstep 4.2 / Prelude to Rhapsody "Easter Egg" > comp.sys.next.programmer "Gregory H. Anderson" <greg@afs.com> writes, > > PS: The reason I cross-posted this to the programmer groups was > > to make the point that you can test multiple interface styles > > and layouts and sizes, even you if you don't own all the supported > > equipment. Name another cross-platform environment where you can > > do _that_. > > Visualworks\Smalltalk from ParcPlace used to support something like > this, about 4 years ago. Still does. In fact I write my own UILookPolicy's in VisualWorks. I've been waiting for years to be able to do this under OPENSTEP. Look's like it just might happen with the rumored "Theme's API". Ron rpomeroy@dallas.net
From: Jonathan Hendry <jon@subsequent.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 14 Sep 1997 00:16:55 GMT Organization: Netcom Sender: Jonathan Hendry <jon@aahz.ix.netcom.com> Message-ID: <5vfadn$aau@sjx-ixn4.ix.netcom.com> References: <341A2776.150A@afs.com> <5veg35$12e$3@ha2.rdc1.nj.home.com> <5vf62i$iju$1@sps1.phys.vt.edu> Nathan Urban <nurban@sps1.phys.vt.edu> wrote: > The NSInterfaceStyle default also works under 4.1. Try it! Does it change the menu style? On 4.2, the menus remained NeXT-style, but other UI elements did change. (The ProjectBuilder browser got an atrocious white background, presumably because I don't have that Windows-UI color set.)
Newsgroups: comp.sys.next.programmer From: ernst@fritz.Xsnafu.de (Ernst Kloecker) Subject: Re: Wait Cursor on NT Message-ID: <EGExnL.5vp@fritz.snafu.de> Sender: news@fritz.snafu.de Organization: Software Engineering References: <5v8vc3$eb4$1@concorde.ctp.com> <5varvp$fjr@mochi.lava.net> <5vb5jk$7j6$1@concorde.ctp.com> Date: Fri, 12 Sep 1997 20:45:17 GMT In article <5vb5jk$7j6$1@concorde.ctp.com>, dcoyle@ctp.com (David A. Coyle) wrote: >arti@lava.DOTnet (Art Isbell - remove "DOT") wrote: > >>#import <System/Windows.h> >>#import <System/Winnt.h> >> >> SetCursor(LoadCursor(NULL, IDC_WAIT)) >> >> Doesn't this make you just want to get into more Win32 programming ?-) >>Looks like a lotta fun. "NT" stands for "New Technology", right? > >Yep. ;-) > >Now, how do I turn it off? :-) > HCURSOR oldCursor; oldCursor = SetCursor(LoadCursor(NULL, IDC_WAIT)); . . . SetCursor (oldCursor); (Please remove 'X' in email address to reply) --- Ernst Kloecker ernst@fritz.Xsnafu.de
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer Subject: Re: Replacing hard disk on Intel NS 3.2 Followup-To: comp.sys.next.sysadmin Date: 14 Sep 1997 01:57:34 GMT Organization: none Message-ID: <5vfgae$12e$7@ha2.rdc1.nj.home.com> References: <341AC34E.159E@decide.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: bobk@decide.com NOTE: FOLLOWUPS TO COMP.SYS.NEXT.SYSADMIN In <341AC34E.159E@decide.com> Bob Korsan wrote: > Can anyone lay out the steps that I would have to follow? Well, the important thing is that you make sure you have a copy of all the drivers which you are using. (I suggest running Configure.app and taking a TIFF of them using Grab.app so you have something to refer to later when and if you have to reconfigure things). Do you have the 3.2 drivers floppy disk? I think /usr/Devices/ and/or /private/Drivers/i386/ is the place where they get installed. Someone else can verify that. > Can anyone point me to a web site or ftp site which would have the info > I would like? (I tried NeXTanswers, but after about 2-3 hours of > wandering and searching that site, I couldn't find anything. I sent a > message to Apple Enterprise Support, but they say the OS is no longer > supported. They recommended I message this group, so I am.) They suggested csn.programmer? That seems strange. Followups set to csn.sysadmin, since that seems to make more sense. > Lastly, I haven't had any trouble with the OS, so I don't see the need > to upgrade, but if there is something really wiz bang on the latest > version, could someone let me know what I would be getting? And if you > know the price to upgrade from 3.2 to the latest (I don't even know what > the latest version is), please let me know. If you have the edu version, you can't upgrade. The only really worthwhile thing if you are just a user is the Mail.app from 3.3 which can handle bundles and MIME. Other than that 4.1 is largely the same as 3.2 in my experience. TjL -- I'm sick of SPAM. Therefore, my email address is invalid. To email me, use one of these without spaces: luomat + next @ luomat.peak.org (for NeXT related stuff) luomat + peak @ luomat.peak.org (for PEAK related stuff)
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 14 Sep 1997 01:48:09 GMT Organization: none Message-ID: <5vffop$12e$5@ha2.rdc1.nj.home.com> References: <341A2776.150A@afs.com> <5veg35$12e$3@ha2.rdc1.nj.home.com> <5vf62i$iju$1@sps1.phys.vt.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: nurban@sps1.phys.vt.edu In <5vf62i$iju$1@sps1.phys.vt.edu> Nathan Urban wrote: > In article <5veg35$12e$3@ha2.rdc1.nj.home.com>, nospam@all.please (Timothy J. Luoma) wrote: > > > In <341A2776.150A@afs.com> "Gregory H. Anderson" wrote: > > > > Sure looks like the beginnings of an Appearance Manager to me. But what > > > do I know, I'm just a simple country programmer... 8^) > > > Since I'm stuck with lowly 4.1 rather than 4.2, I'm not sure what this means. > > The NSInterfaceStyle default also works under 4.1. Try it! But I don't know what difference I would be looking for.... 4.1 looks like 3.3 AFAICT It is just a programmer's thing? TjL -- I'm sick of SPAM. Therefore, my email address is invalid. To email me, use one of these without spaces: luomat + next @ luomat.peak.org (for NeXT related stuff) luomat + peak @ luomat.peak.org (for PEAK related stuff)
From: Joseph Panico <jpanico@online.disney.com> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: Mon, 08 Sep 1997 09:56:42 -0400 Organization: Disney Online Message-ID: <3414041A.E71BFBCD@online.disney.com> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > So, does Java have a fragile base class problem? > No it does not. But this is just one of the many broken aspects of Obj-C that Java fixes. IMHO, Java is a far superior language/environment to ObjC. I have been developing, simultaneously, in both languages for almost 2 years now (ObjC for 5). I can tell you that I'm at the point where I dread having to code ObjC. Here is what ObjC can do that Java cannot: 1) Categories-- A usefull tool for cleaning up designs. Without them, you end up with a lot of "helper" classes-- methods that operate on data residing in other classes. Obviously, this is contrary to the spirit of OOP, is error prone and design messy. 2) Forwarding-- Forwarding is usefull for things like proxy classes and seemless method invocation failover. It's an elegant design pattern that can go a long way with a little creativity. Having it can conceptually simplify designs. Here is what Java can do that ObjC cannot: 1) Top to bottom Unicode support. Unicode is built into the foundations. You can program in Unicode. 2) Garbage collect. No more explicity memory management. NeXt's reference counting helps, but it is still error prone. This *vastly* increases the robustness of Java over ObjC 3) Macine independent bytecodes. No more recompilation for platforms. You no longer have to worry about having the right tyhpe of dynamic library for the right platform. Is it a shlib, is it a .dll? Is it installed in the right place? Who cares! This nonsense has been greatly simplified with platform independent bytecodes combined with dynamic class loading based on a CLASSPATH. 4) Objects only! aside from a small number of primitive classes, Java is a pure OOP. No more legacy c stuff- structs pointers, etc. This vastly increases the robustness of Java over ObjC. It also conceptually simplifies the language. 5) Package(ing) of name space. Packages elegantly solve ObjC's irritating problem of having a global Object name space. No more worries of name collision and no more nonsensical naming conventions. 6) Dynamic class loading! This is a huge win! No more minute plus link phases. This is why I dread programming ObjC. The combination of the compiler and linker are about 100 times slower than the combination of compile/load using Symantec Cafe. The current crop of Java environments (Cafe, Asymmetrix, VisualAge) are so fast, that you are essentially working in an interpreted environment. 7) Structured exception hadling that is tightly integrated into the languaged, not hacked on after the fact as a series of macros. As a result, Java's exception handling works better than ObjC's 8) Concurrent programming built into the core of the language. The root class Object knows about concurrent programming constructs. 9) Multi vendor support. Zillions of vendors are competing to produce Java tools. ObjC comes from a single source, and a very flaky one at that. ...there are many others. I just touched on the ones that come to mind right now. The base language of Java is a total no-brainer compared with ObjC. Using java over ObjC should be a layup...but... the only problem that Java currently suffers is that the class libraries are immature compared with most other platforms (especially Next). There is no foundation (yet) and there is nothing as powerfull and elegant as EOF. Fortunately this will be solved within the next 6 months, at which point you should take the layup ;) good luck. > There has been lots of discussion about the differences between Java > and > C++, but I don't think I've seen much about the differences between > Java > and Objective-C. > > Can anyone offer a list of Objective-C features that an Objective-C > programmer would miss if he were forced to use Java? > > For instance, Java doesn't have anything akin to Objective-C's > ForwardInvocation method, does it? > -- > Robert Fisher > Renegade Software > rfisher@onr.com
From: nurban@sps1.phys.vt.edu (Nathan Urban) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 14 Sep 1997 00:09:28 -0400 Organization: Data Systems Consulting, Inc. Message-ID: <5vfo1o$k4a$1@sps1.phys.vt.edu> References: <341A2776.150A@afs.com> <5veg35$12e$3@ha2.rdc1.nj.home.com> <5vf62i$iju$1@sps1.phys.vt.edu> <5vfadn$aau@sjx-ixn4.ix.netcom.com> In article <5vfadn$aau@sjx-ixn4.ix.netcom.com>, Jonathan Hendry <jon@subsequent.com> wrote: > Nathan Urban <nurban@sps1.phys.vt.edu> wrote: > > The NSInterfaceStyle default also works under 4.1. Try it! > Does it change the menu style? No. That's the main reason why there are two NIBs in IB..
From: sanguish@digifix.com Newsgroups: comp.sys.next.advocacy,comp.sys.next.announce,comp.sys.next.hardware,comp.sys.next.software,comp.sys.next.misc,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.sys.next.programmer Subject: NEXTSTEP/OpenStep Resources on the Net Supersedes: <8835873604833@digifix.com> Date: 14 Sep 1997 03:49:15 GMT Organization: Digital Fix Development Message-ID: <9488874209621@digifix.com> Topics include: Major OpenStep/NEXTSTEP World Wide Web Sites OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups Major OpenStep/NEXTSTEP FTP sites NeXTanswers Major OpenStep/NEXTSTEP World Wide Web Sites ============================================ The following sites are a sample of the OpenStep related WWW sites available. A comprehensive list is available on Stepwise. Stepwise OpenStep/Rhapsody Information Server http://www.stepwise.com Stepwise has been serving the OpenStep/NEXTSTEP community since March 1993. Some of the many resources on the site include: OpenStep Third Party Software guide, Developer Directory, Mailing List information, extensive listing of FTP and WWW sites related to OpenStep and NEXTSTEP, OpenStep related Frequently Asked Questions. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. This is the World Wide Web interace to the FTP site. Apple Enterprise Software Group (formerly NeXT Computer, Inc.) http://www.next.com Here is where you'll find the NeXTanswers archive, with information on OpenStep installation, drivers and software patches. Apple Computer's 'Prelude to Rhapsody' Self Support Site http://devworld.apple.com/dev/prelude.html This site has been constructed to help you help yourself to learn as much as possible about the foundation for Rhapsody, today's OPENSTEP. The site provides an informal collection of pointers, references, and starting points for developers who are using the Prelude to Rhapsody bundle, distributed at this year's Worldwide Developer Conference. OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups ==================================================== COMP.SYS.NEXT.ADVOCACY This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. COMP.SYS.NEXT.ANNOUNCE Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. COMP.SYS.NEXT.BUGS A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT-specific groups as well. COMP.SYS.NEXT.HARDWARE Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. COMP.SYS.NEXT.MARKETPLACE NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. COMP.SYS.NEXT.MISC For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! COMP.SYS.NEXT.PROGRAMMER Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. COMP.SYS.NEXT.SOFTWARE This is a place to talk about [third party] software products that run on NEXTSTEP systems. COMP.SYS.NEXT.SYSADMIN Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. ** RELATED NEWSGROUPS ** COMP.SOFT-SYS.NEXTSTEP Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. COMP.LANG.OBJECTIVE-C Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. COMP.OBJECT Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep next-advocacy next-announce next-bugs next-hardware next-marketplace next-misc next-programmer next-software next-sysadmin object lang-objective-c (For a full description, send mail to listserv@antigone.com). The subscription syntax is essentially the same as Majordomo's. To subscribe, send a message to *-request@lists.best.com saying: subscribe where * is the name of the list e.g. next-programmer-request@lists.best.com Major OpenStep/NEXTSTEP FTP sites ================================= ftp://ftp.next.peak.org The main site for North American submissions formerly ftp.cs.orst.edu ftp://ftp.peanuts.org: (Peanuts) Located in Germany. Comprehensive archive site. Very well maintained. ftp://ftp.nluug.nl/pub/comp/next NeGeN/NiNe (NEXTSTEP Gebruikers Nederland/NeXTSTEP in the Netherlands) ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! _________________________________________________________________ Written by: Eric P. Scott ( eps@toaster.SFSU.EDU ) and Scott Anguish ( sanguish@digifix.com ) Additions from: Greg Anderson ( Greg_Anderson@afs.com ) Michael Pizolato ( alf@epix.net ) Dan Grillo ( dan_grillo@next.com )
From: Jonathan Hendry <jon@subsequent.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 14 Sep 1997 04:24:56 GMT Organization: Netcom Sender: Jonathan Hendry <jon@aahz.ix.netcom.com> Message-ID: <5vfouo$5re@sjx-ixn8.ix.netcom.com> References: <341A2776.150A@afs.com> <5veg35$12e$3@ha2.rdc1.nj.home.com> <5vf62i$iju$1@sps1.phys.vt.edu> <5vfadn$aau@sjx-ixn4.ix.netcom.com> <5vfo1o$k4a$1@sps1.phys.vt.edu> Nathan Urban <nurban@sps1.phys.vt.edu> wrote: > In article <5vfadn$aau@sjx-ixn4.ix.netcom.com>, Jonathan Hendry <jon@subsequent.com> wrote: > > Nathan Urban <nurban@sps1.phys.vt.edu> wrote: > > > The NSInterfaceStyle default also works under 4.1. Try it! > > Does it change the menu style? > No. That's the main reason why there are two NIBs in IB.. Sure. I'd assumed that, when NSInterfaceStyle is set, the app would use the appropriate nib if available. It didn't on my machine, and I was wondering if it should have, or if other people are seeing that behavior.
From: Steve Kellener <skellener@earthlink.net> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: Sat, 13 Sep 1997 23:09:42 -0700 Organization: EarthLink Network, Inc. Message-ID: <341B7F4C.5EEA@earthlink.net> References: <341A2776.150A@afs.com> <5veg35$12e$3@ha2.rdc1.nj.home.com> <5vf62i$iju$1@sps1.phys.vt.edu> <5vfadn$aau@sjx-ixn4.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Jonathan Hendry <jon@subsequent.com> Jonathan Hendry wrote: > > Nathan Urban <nurban@sps1.phys.vt.edu> wrote: > > > The NSInterfaceStyle default also works under 4.1. Try it! > > Does it change the menu style? On 4.2, the menus remained NeXT-style, > but other UI elements did change. I think the The NSInterfaceStyle default is intended to allow NT users to see and use a NeXT interface. It wouldn't be necessary to set it on Mach.
From: Jim Redman <jim@ergotech.com> Newsgroups: comp.sys.next.programmer Subject: Re: OpenStep Linking error on bundle Date: Sun, 14 Sep 1997 04:51:05 -0600 Organization: ErgoTech Message-ID: <341BC199.7713@ergotech.com> References: <34142B50.41C6@ina.fr> <5v1eai$lni@shelob.afs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Gregory H. Anderson wrote: > The problem is that the NT linker does not observe the -ObjC or > -include_all linker symbols. Therefore, it only brings in classes to > which there are explicit references elsewhere in your source code. > (This bit me a few times with classes that only exist in NIB files.) > > The solution is to refer to the missing class explicitly in some > piece of code that you know will be included. I usually put a harmless > [ClassA class] statement in an +initialize block. That simple reference > will force the linker to resolve the ClassA symbol. This problem does not exist with dynamically loaded bundles or frameworks, at least in the case described where a superclass used in the bundle or framework is unresolved. I'm not sure that I have tested other cases. In these cases the framework or bundle will load correctly although you probably need to force the load with a request for the principal class. This problem does exist with frameworks that are linked directly against the application. Jim
From: pecora@zoltar.nrl.navy.mil (Lou Pecora) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Sun, 14 Sep 1997 07:46:37 -0500 Organization: Naval Research Lab Message-ID: <pecora-1409970746370001@esp225.nrl.navy.mil> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <5upjgt$sme$1@news.lth.se> <LTaylor7-0509972346490001@vic-ca1-24.ix.netcom.com> <5uu4ai$47v$1@news.lth.se> <LTaylor7-0709970614470001@vic-ca1-03.ix.netcom.com> <5uur7i$9po$1@news.lth.se> <19970907234113405260@accs-as15-dp07.dlls.grid.net> <jonwhite-1309971756090001@phd-as2s21.erols.com> In article <jonwhite-1309971756090001@phd-as2s21.erols.com>, jonwhite@erols.com (J. White) wrote: > Your absolutely right. Perception and marketing sells products. Apple > has always been pathetic with their marketing. And now the perception > factor is against them in IS departments. My company uses mac office > machines and unix boxes to run our business. With Apple's recent moves, > now I must ask myself if I'm comfortable buying Mac systems, possibly > jeopardizing our companies competitiveness. I've decided that this > monday, I'll call Apple and ask them why I should keep buying their > machines. There answer had better be very good. Please post the answer. Lou Pecora code 6343 Naval Research Lab Washington DC 20375 USA == My views are not those of the U.S. Navy. == == No spaming or solicitations -- both are illegal at this site. ------------------------------------------------------------ The 4th Experimental Chaos Conference Home Page: *** The conference is over, but see what you missed: *** Titles, and Abstracts online http://natasha.umsl.edu/Exp_Chaos4/ ------------------------------------------------------------- Next Experimental Chaos Conf. in Torino, Italy 1999 !!! ------------------------------------------------------------
From: seanlNoSpam@carmi.cs.umd.edu (Sean Luke) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 14 Sep 1997 17:02:41 GMT Organization: University of Maryland at College Park Message-ID: <5vh5bh$79q$1@walter.cs.umd.edu> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> <3414041A.E71BFBCD@online.disney.com> Joseph Panico (jpanico@online.disney.com) wrote: >Here is what ObjC can do that Java cannot: > >2) Forwarding-- Forwarding is usefull for things like proxy classes and >seemless method invocation failover. It's an elegant design pattern that >can go a long way with a little creativity. Having it can conceptually >simplify designs. While I'm also a big fan of Java, this point CANNOT be stressed heavily enough. Java does not have sufficient guts to do distributed objects in anything approaching the elegant way that Objective-C does (course, garbage collection in distributed objects would also be a mess, but that's another tale). In my opinion, this is a serious failing for a language that claims to be truly dynamic. Sean Luke seanl@cs.umd.edu
From: tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C?... categories contra OOP ?? Date: 14 Sep 1997 20:10:35 GMT Organization: Cambridge Technology Partners, Inc. Message-ID: <5vhgbr$c92$1@concorde.ctp.com> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> <3414041A.E71BFBCD@online.disney.com> Joseph Panico <jpanico@online.disney.com> wrote: >Here is what ObjC can do that Java cannot: > >1) Categories-- A usefull tool for cleaning up designs. Without them, >you end up with a lot of "helper" classes-- methods that operate on data >residing in other classes. Obviously, this is contrary to the spirit of >OOP, is error prone and design messy. > Maybe I'm missreading this but are you saying that _categories_ violate the OOP concept ? Since categories usually are just shortcuts for a combiantion of other methods they IMHo fit very nicely into an OO world and make life a lot lot simpler. Even in reality most things are done with "categories" and therefor the logical concept transfers quite nicely and is easy to understand as far as I can tell. Error prone and desing messy might be valid in some cases..but in other its not. I gues this could be said of almost everything. You can use a tool in a way which it was not inteded for. Aloha Tomi P.S. Btw.. what I would like to play around with are "context-sensitive" methods. Imagine an object implements two protocols which both include the "cookMeal" method. But if the <NSBehaveLikeBeingAtHome> context is the objects default it would order a pizza....and if the <NSYouAreAtWorkChef> context is specified the cookMeal would check for the next meal on the chefs order list. To some degree this is kind of morhping the objects type from "person" to "chef"...but still it is exaclty the same object. Having a fix for that would help if an object has to implement two protocols/interfaces which happen to conflict at certain methods...but in the code it is always clear which context the object currently is used in.
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <341b1b11.0@news.west-tech.com> Control: cancel <341b1b11.0@news.west-tech.com> Date: 14 Sep 1997 22:53:08 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.341b1b11.0@news.west-tech.com> Sender: info@pgrs.com Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: nurban@sps1.phys.vt.edu (Nathan Urban) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C?... categories contra OOP ?? Date: 14 Sep 1997 19:06:48 -0400 Organization: Data Systems Consulting, Inc. Message-ID: <5vhqm8$phn$1@sps1.phys.vt.edu> References: <01bcb30b$32787ae0$0100000a@ivanova> <01bcb4cf$9b253c70$0100000a@ivanova> <3414041A.E71BFBCD@online.disney.com> <5vhgbr$c92$1@concorde.ctp.com> In article <5vhgbr$c92$1@concorde.ctp.com>, tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) wrote: > Joseph Panico <jpanico@online.disney.com> wrote: > >Here is what ObjC can do that Java cannot: > >1) Categories-- A usefull tool for cleaning up designs. Without them, > >you end up with a lot of "helper" classes-- methods that operate on data > >residing in other classes. Obviously, this is contrary to the spirit of > >OOP, is error prone and design messy. > Maybe I'm missreading this but are you saying that _categories_ violate the > OOP concept ? It sounds like he was saying that having lots of helper classes is contrary to the spirit of OOP, etc.
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy,comp.sys.next.programmer Subject: How did Semper.fi die? (was Re: Apple developers continue to shift to Intel, MOT to sue? Date: 14 Sep 1997 18:39:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B041E51A-926E6@206.165.44.29> References: <199709141729391030613@accs-as31-dp14.dlls.grid.net> To: "AIMED-Talk" <aimed-talk@aimed.org>, "John Bauer" <drifterusa@macconnect.com> nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.next.programmer MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit John Bauer <drifterusa@macconnectNOT.com> said: > Lawson English wrote: > > > You can't ask it on Semper.fi, the official Macintosh developer feedback > > list that was created by Guy Kawasaki, because Semper.fi is no longer > > active. > > Lawson, I'm curious to hear the story of the demise of Semper.fi (not > being a developer). Care to spin the tale? My version: Well, it all goes back to GX... ;-) Actually, it does, in a way. When the NeXT purchase was announced, some of us GX supporters (me, mostly) waged a PR battle against DPS-only and for GX in Rhapsody on Semper.fi. I made a few choice remarks about Amelio's (Jobs, it turns out) decisions, suggesting that Amelio was raping APple of its technologies, and that only a tiny handful of the technologies would survive -mostly within QTML (turns out that I was correct, although it appears to have been Jobs behind it all and not Amelio). Lots of folks took offense at the rapist remark. Others took offense at what they saw as the monopolization of the list by myself and a few others. Others, I truely believe, felt threatened that even one single NeXT technology might not be used as the One True Way. Actually, to be perfectly honest, while I *WAS* singled out as the Bad Boy of semper.fi, I believe that it had become too much of a lightning rod for all developers to publicly express their negative opinions about Apple's new policies. Apple (Jobs) didn't want to hear or have the press (who often lurked on semper.fi) hear just how unpopular Apple policy was becoming with the developers. They tried moderation, but that wasn't satisfactory. They now have... Nothing. No open forum to provide official feedback to Apple. Nada. [and its all MY fault <sniff>] OF course, if you, or anyone else, wants to deal with developers in an unmoderated, open-to-the-public forum, there's always AIMED-talk. <http://www.aimed.org> AIMED = Association of INDEPENDENT Macintosh Engineers and Developers ------------------------------------------------------------------------- "Apple's the Princess Diana of the computer industry." -Larry Campbell "And Steve Jobs has taken over for the regular driver." -Takechan <take@oatmeal.com> -------------------------------------------------------------------------
From: leigh@cs.uwa.edu.au (Leigh Smith) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 15 Sep 1997 02:49:10 GMT Organization: The University of Western Australia Message-ID: <5vi7n6$6fp$1@enyo.uwa.edu.au> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> <3414041A.E71BFBCD@online.disney.com> <5vh5bh$79q$1@walter.cs.umd.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: seanlNoSpam@carmi.cs.umd.edu Without trying to turn this into a religous war, keyname parameters of Objective C and avoiding the use of static to nonintuitively mean class method or variables are *really* nice features of ObjC. Sure they are syntactical issues and any programmer worth their salt will not be impeded by them, but given Java was developed from direct exposure to ObjC and the fact that Sun is marketing it to become the standard we all have to put up with for god knows how many more years, couldn't they have stolen those ideas as well? -- Leigh Computer Science, University of Western Australia Smith +61-8-9380-3778 leigh@cs.uwa.edu.au (NeXTMail/MIME) "Home pages are the pet rock of the 90s. We all have them, we all think they're very cute. But in a few years we're going to look back and be pretty embarrassed." -- Tony Shepps <toad@pond.com> "Why wait?" -- Peter Langston
Newsgroups: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: How did Semper.fi die? (was Re: Apple developers continue to shift to Intel, MOT to sue? References: <199709141729391030613@accs-as31-dp14.dlls.grid.net> <B041E51A-926E6@206.165.44.29> From: John Kheit <jkheit@xtdl.com> Organization: monoChrome, Inc., NJ, USA Message-ID: <341cab2e.0@news.xtdl.com> Date: 15 Sep 97 03:27:42 GMT "Lawson English" <english@primenet.com> wrote: > [and its all MY fault <sniff>] I tend to agree with Lawson...if the postings of a single person were somehow the "reason" for it dying...it's likely it wasn't all to healthy a venture to start. Besides, everyone knows all the good stuff is in usenet; BARRING THE CRAZY CROSS POSTING...Lawson, you really have to calm down with it...especially all the programming groups...csn.programming is looking more like a tabloid press area than a programming area...maybe c.s.mac.programming is supposed to be that way (dunno, I dont read it or its charter), but this kind of gossipy stuff doesn't belong there...followups to the APPROPRIATE groups please...) -- Thanks, be well, take care, later, John Kheit; Self expressed... __________________________________________________________________ monoChrome, Inc. ASCII, MIME, PGP, SUN, & NeXTmail OK NeXT/OPENSTEP Developer mailto:jkheit@xtdl.com Telepathy, It's coming... http://www.xtdl.com/~jkheit Franklin Pierce Law Center You're dangerous because you're honest the rest of us back in reality might consider different and better avenues... -- Thanks, be well, take care, later, John Kheit; Self expressed... __________________________________________________________________ monoChrome, Inc. ASCII, MIME, PGP, SUN, & NeXTmail OK NeXT/OPENSTEP Developer mailto:jkheit@xtdl.com Telepathy, It's coming... http://www.xtdl.com/~jkheit Franklin Pierce Law Center You're dangerous because you're honest
From: bwanga@noUCE.cats.ucsc.edu (Timothy A. Seufert) Newsgroups: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: How did Semper.fi die? (was Re: Apple developers continue to shift to Intel, MOT to sue? Date: Sun, 14 Sep 1997 23:00:21 -0700 Organization: Evil Geniuses For A Better Tomorrow, Inc. Message-ID: <bwanga-1409972300220001@tsb-33.ucsc.edu> References: <199709141729391030613@accs-as31-dp14.dlls.grid.net> <B041E51A-926E6@206.165.44.29> In article <B041E51A-926E6@206.165.44.29>, "Lawson English" <english@primenet.com> wrote: [snip extremely slanted and self-important view of why Semper.fi "died", complete with unsupported assertions about other matters] Lawson, sometimes you really astound me with your inability to accept that not everything Apple does is part of some Evil Plan. Did you actually read Chuq von Rospach's list of reasons why Semper.fi should be put on hold? It sure doesn't sound like it. (For those who aren't aware, Chuq runs Apple's various mailing lists. He was the one who recently made the decision to put semper.fi on indefinite hiatus.) A sampler of what I remember (I don't have the original mailing to make sure I'm accurately representing Chuq, but at least I'm trying, unlike you): 1. Without moderation, semper.fi quickly degenerates into a total flame-fest. Just as bad as what goes on here. The signal-to-noise ratio is horrible. Not just Chuq's opinion; he requested feedback from subscribers and (from memory) over 95% of it was negative about the list's content. 2. Chuq took on the position of moderator for the sake of convenience, but it is actually not his place to be the moderator of an Apple Developer Relations mailing list. It's not part of his job description, he doesn't have time for it, and he doesn't have the right to effectively speak for ADR via his moderation. 3. Nobody else seems to want to moderate it. 4. Chuq analyzed the change in demographics over time, and it is evident that few people at Apple are subscribed to semper.fi these days. The number has been declining steadily. This is not surprising, because even with moderation semper.fi doesn't do much of a job of communicating valid concerns to Apple. Most posters fail to really think through their opinions, so what they post to it does not amount to anything like a good analysis of why a particular thing is the wrong thing for Apple to do (which is the kind of feedback one presumes Apple needs). But it does a heck of a job of being pretty hostile to their employees, driving them away from the list. Therefore, semper.fi no longer serves as an effective communications channel to Apple. 5. Although semper.fi is theoretically supposed to be a way for developers to communicate with apple, theory is different from practice. Anybody, developer or not, can pound the pulpit on semper.fi. Furthermore, there is no method of enforcing that Apple employees listen... there is nobody in ADR appointed to be The Guy Who Watches Semper.fi For Important Stuff, because semper.fi was not and never has been an *official* way to get in touch with Apple. More support for 4. I believe there was more... but hopefully by now I've provided enough of the real reasons for semper.fi's vacation to show that you are (as usual) talking through your ass. Despite your heavy implication that Steve Jobs shut it down because he wants people silenced, it was Chuq's decision, and he made it because there were severe problems with the list. Oh, by the way, folks, Chuq isn't shutting it down for all time... he just doesn't want to start it up again until he or somebody else can come up with a non-band-aid approach to making it a worthwhile resource. Moderation was the attempted band-aid, and it failed. -- Tim To mail me unsolicited advertisements: Move to Siberia. Wait until I say it's OK to send. Everybody else, remove "noUCE." from my address.
From: pemmerik@solair1.inter.NL.net (P.J.L.van Emmerik) Newsgroups: comp.sys.next.programmer Subject: Memory leak when using Enterprise Objects Framework 1.1 Date: Mon, 15 Sep 1997 07:14:50 GMT Organization: NLnet Message-ID: <5vin9d$j0s$1@news.NL.net> I am using Enterprise Object Framework 1.1 to access an Oracle database. There are some tables that, when displayed on screen, i always want to display the 'current' state of the database. I have tryd to sove this by doin a fetch from a timed entry. When i do this i get a memory leak i cannot find. I have tryd to use my own autorelease pool in the method that handles the timed entry. I hav alse tryd the 2082_EOApplicationPatch.o patch. Both didn't help. Any suggestions? Please Email to: emmerik@hpb.holec-projects.nl P.J.L. van Emmerik Holec Projects B.V. Email: emmerik@hpb.holec-projects.nl PO.BOX 565, 7550 AN Hengelo pemmerik@solair1.inter.NL.net The Netherlands Phone: +31 74 2558 688 --
From: scott@doubleu.com (Scott Hess) Newsgroups: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: How did Semper.fi die? (was Re: Apple developers continue to shift to Intel, MOT to sue? Date: 15 Sep 1997 07:05:45 GMT Organization: Is a sign of weakness Message-ID: <SCOTT.97Sep15024112@howard.doubleu.com> References: <199709141729391030613@accs-as31-dp14.dlls.grid.net> <B041E51A-926E6@206.165.44.29> In-reply-to: "Lawson English"'s message of 14 Sep 1997 18:39:00 -0700 In article <B041E51A-926E6@206.165.44.29>, "Lawson English" <english@primenet.com> writes: Actually, to be perfectly honest, while I *WAS* singled out as the Bad Boy of semper.fi, I believe that it had become too much of a lightning rod for all developers to publicly express their negative opinions about Apple's new policies. Apple (Jobs) didn't want to hear or have the press (who often lurked on semper.fi) hear just how unpopular Apple policy was becoming with the developers. They tried moderation, but that wasn't satisfactory. They now have... Nothing. No open forum to provide official feedback to Apple. Nada. [and its all MY fault <sniff>] Well, I think you can afford some modesty in this particular case. NeXT has never been much for constructive criticism. For instance, one might say something polite and minor about a strange keyboard decision (someone decided to move the \ and | keys, important to C and shell use, from the main area of the keyboard off to E.T. finger land) on a private list, only later to find out that the list was actually forwarded to relatively higher-ups at NeXT, later to be labelled behind your back as someone who's "hard to work with", and have clients say to you "You aren't nearly as impossible to work with as my NeXT rep lead me to believe." [Well, now that I've got _that_ off my chest...] In any case, NeXT has always had this weird way of approaching criticism. First, the criticism is denied ("We've considered that, but it's not a problem because you shouldn't be using that technique."). Then, the criticizer is denied ("We won't address that issue, obviously someone has an axe to grind."). Then they flip-flop, and take the original criticism and the original suggestions, blow them _entirely_ out of proportion, and make that their new goal in life ("Per customer request, we'll be dropping hardware now."). [Sorry, that last needs more interpretation. Customers _were_ requesting more hardware alternatives, namely Intel PCs. They _weren't_ requesting that the existing alternatives be dropped. One only had to look at the chaos that particular decision created to see how much customer consultation went into it ...] Of course, that's _also_ been the way Apple has always approached these things. I think it's something to do with persecution complexes. If everyone is after you, are you paranoid or just sensible? Depends on if you spend your time taking out those who are _really_ after you - or if your time is spent sniping at your friends, who are just trying to point you in the right direction. -- scott hess <scott@doubleu.com> (606) 578-0412 http://www.doubleu.com/ <Favorite unused computer book title: The Demystified Idiots Guide to the Zen of Dummies in a Nutshell in Seven Days, Unleashed>
From: mjm@attila.apana.org.au (Matt McLeod) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: 15 Sep 1997 18:58:36 +1000 Organization: APANA Hunter Regional Hub Message-ID: <5vitbs$3vu$1@attila.apana.org.au> References: <5urae7$f7c$1@attila.apana.org.au> <alex-0809970817390001@p1-40.van.tvs.net> <5v2qsf$dkv$1@attila.apana.org.au> <341c133c.0@news.xtdl.com> In article <341c133c.0@news.xtdl.com>, John Kheit <jkheit@xtdl.com> wrote: :mjm@attila.apana.org.au (Matt McLeod) wrote: :> (a) Make the software free. Give away the source. Put Rhapsody :> under GPL (b) Make the development tools free too. (c) Recruit :> programmers who are actually interested in the project. Not :> just people who want a job. :> (d) Keep management away from the programmers. Set some basic :> goals, :> then get the hell out. : :> The company then becomes the maintainer of the software, and :> still sells it as a shrinkwrap product (nothing in the GPL prevents :> this). Make your money on shrinkwrap sales to the consumer :> market, support, and consulting. : :GPL is lunacy incarnate. There are far better models for commercial :ventures...namely MiscKit. Regardless, the idea of apple giving :it away strikes me as lunacy in and of itself. Why don't people :ever ask ms to give away stuff? I guess the more communistic might. :Regardless, the rest of us back in reality might consider different :and better avenues... I don't see the problem with giving stuff away. GPL might not be the right license to use, but it's a reasonable starting-point. Even going part-way would be a good move, IMO. Give away Rhapsody binaries to anyone who wants them. If you want support, or a pretty box, or hardcopy documentation, then you buy it. That at least gets you a lot more users than you'll get if you're selling the thing for US$400. I may be going out on a limb here, but my guess is that Rhapsody/Intel will probably sell about as well as OS/2 does. They'll pick up current OpenStep users, maybe steal some of the OS/2 userbase, tempt a few Unix users across, and maybe some of the more reluctant Windows users. That's with a pricetag of over US$200 or so. Even if the product is excellent, more than US$200 is not going to be an impulse purchase for most, and you need people to give it a try "just for the hell of it" - otherwise you won't get many new users. Reduce the price, and they'll get more "for-the-hell-of-it" purchasers. People who like the idea of an alternate OS, but aren't going to spend big bucks just to try the thing out. Some of them will stick with it. The more users, the bigger the software market. Which is good for Apple (Claris) and for ISVs. Potentially good for hardware sales, too - once your OS and apps all run under Rhapsody/Intel, when it comes to upgrade-time, PowerMacs suddenly become a feasible option. -- Matt McLeod, <mjm(at)attila.apana.org.au> "Please try to understand before one of us dies".
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: How did Semper.fi die? (was Re: Apple developers continue to shift to Intel, MOT to sue? Date: 15 Sep 1997 02:12:00 -0700 Organization: Primenet Services for the Internet Message-ID: <B0424F43-221804@206.165.44.29> References: <bwanga-1409972300220001@tsb-33.ucsc.edu> nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.next.programmer MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Timothy A. Seufert <bwanga@noUCE.cats.ucsc.edu> said: > > 1. Without moderation, semper.fi quickly degenerates into a total > flame-fest. Just as bad as what goes on here. The signal-to-noise ratio > is horrible. Not just Chuq's opinion; he requested feedback from > subscribers and (from memory) over 95% of it was negative about the > list's > content. But WHY does semper.fi degenerate so quickly? It wasn't that way when it first started, was it? At one time, semper.fi (set up by Guy Kawasaki as part of his role as Chief Evangelist of Apple -an Apple Fellow, no less) DID provide useful feedback without moderation. Now, one can say that it WAS my fault (I've gotten e-mail to that effect) or one can say that it was NOT my fault (and that I've got ego problems), but the fact remains that semper.fi DID degenerate. The "why" question, I believe, is answered by answering the "when" question: When did semper.fi start to degenerate? I believe that Semper.fi started to degenerate after Hancock pulled the plug on Copland DR1 and no word was given for many months as to what Apple would do "next." This frustrated MANY developers. The FINAL degeneration started when NeXT was purchased and ALL of Apple's Mac developers were apparently abandoned/marginalized. Heidi Roizen resigned and her replacement was not introduced to the list until a week later, even though he had been named when Roizen resigned. GX was crippled/canned, even in MacOS, without a clear path for developers to handle printing. OpenDoc and a host of other technologies were killed/put on life-support during the Ides of March shootemup. Developers didn't know what to expect next. In a nutshell: technologies that Apple had touted as the "next big thing" for years, that developers had spent months and years learning about and implementing end-user solutions with, were canned or apparently canned. Massive chaos reigned at Apple and developers were left clueless. For instance, the original March 14 announcement concerning Game Sprockets implied that Yellow BOx would NOT have games support and that all games would have to be done within BLue Box (this was so repugnant that a correction/spin-doctoring was done on the comp.sys.mac.programmer.games newsgroup the same day as the initial March 14 announcement to "clarifiy" (correct/contradict) the original announcement -the original announcement was even pulled off of Apple's web-site and the correction inserted as though it were the original). The list goes on and on of the many instances of _faux pas_ that Apple has commited as far as developer-relations goes that semper.fi provided feedback about. To expect semper.fi -the official (unofficial) feedback list for Apple developers- to NOT become noisy and rambunctious under such conditions is clearly unrealistic. Now, we have the clone-licensing issue. Mac users may not know what Mac developers are thinking, but it isn't pretty, from what I've seen (and thought myself -contrary to popular opinion, I've been a "professional" mac developer since 1987 when I help provide software support for a 68020 accelerator board for the Mac Plus). There's no longer official (unofficial) feedback list for developers to express their opinions and/or ideas. I no longer see that much sense-of-community in AIMED-talk, or in the csmp.* newsgroups. How can you expect semper.fi NOT to reflect the resentment and hostility of Mac developers in this situation? Rather than close down the list, Mac management and executives and the BoD should be using it to determine just what can be done to keep Mac developers interested in MacOS and Rhapsody. That was the original purpose of the list, as I understand it, and simply saying "it didn't work" ignores WHY it didn't work, and makes it impossible (in my mind) for the "why" to be corrected: if Mac managemnet and executives and the BoD refuse to listen to Mac developers because they are too noisy, then they are dooming themselves to having fewer and fewer Mac developers to listen to. Conversely, you can suggest that it was "all my fault" and that if it weren't for me and a few others like me, semper.fi would still be around, as some e-mail messages have claimed. ------------------------------------------------------------------------- "Apple's the Princess Diana of the computer industry." -Larry Campbell "And Steve Jobs has taken over for the regular driver." -Takechan <take@oatmeal.com> -------------------------------------------------------------------------
From: MARMIER Raphael <marmier4@hei.unige.ch> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software Subject: Re: HELP FDISK disaster!!!!!!!!!! Date: Mon, 15 Sep 1997 15:35:18 +0200 Organization: University of Geneva Message-ID: <Pine.SOL.3.96.970915153258.17593A-100000@hei.unige.ch> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII To: Patrick McGowan <patmvcr@rogers.wave.ca> Hi > Are these partitions lost forever? Is there any way to recover them? I had > backed up most of the files a few weeks ago, but a few important files were > not. I have not reformatted those sectors of the drive or done anything > else since......except look at them in Partition Magic, an interactive > partitioning application in DOS. Can I use something from Norton Disk > Doctor to find the beginning of my old partitions?...I had a lot af spare > space between the original "free" space and my important DOD > partitions...it was set aside as swap space for Windows 95. > > HELP!!!!! If you can find a way to know the begin and end address of your extended partition, you can update the partition table with Norton Disk editor (be carefull with this tool!). This should work as I suppose only the partition table is damaged. Unfortunatly, I don't know the exact syntax... remember it is a Dos extended partition. If there was more damage (ex: to the file system), I think you may be able to repair it with a disk utility after you recover the partition. But before doing so, make absolutely sure you have the right coordinates for your partition! , unless you want to trash it for good :)) good luck cheers PS: This post rather belongs to comp.sys.next.sysadmin ----------------------------------------------------------------------- Raphael Marmier --> marmier4@hei.unige.ch --> http://heiwww.unige.ch/~marmier4 -----------------------------------------------------------------------
From: "Robert A. Decker" <comrade@umich.edu> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 15 Sep 97 12:32:04 -0400 Organization: University of Michigan ITD News Server Message-ID: <B042DB4F-78B2@141.214.128.36> References: <3414041A.E71BFBCD@online.disney.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit On Mon, Sep 8, 1997 9:56 AM, Joseph Panico <mailto:jpanico@online.disney.com> wrote: >Here is what ObjC can do that Java cannot: You forgot one of the most important things: Not having to worry about type in Obj-C. After programming in Obj-C and Newtonscript and a little Dylan I absolutely cannot stand the static typing of Java/C++. In fact, static typing ruins the concept of OO programming. Why should I care what type the object is that I just got? I should just be able to tell it to do something and leave it at that. (that being said, I usually use static typing in Obj-C except for a few functions). rob -- <mailto: "Robert A. Decker" comrade@umich.edu> <http://hmrl.cancer.med.umich.edu/Rob/index.ssi> Programmer Analyst - Health Media Research Lab University of Michigan Comprehensive Cancer Center "Get A Life" quote #18: "Well, you're always saying I can do anything I want to do if I put my mind to it. This is something I can do without putting my mind to it at all." -Chris Elliott in "The Prettiest Week of My Life"
From: bettis@inetnebr.com (Mr. Jeremy Bettis) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 15 Sep 1997 12:33:03 -0500 Organization: Internet Nebraska Message-ID: <5vjrgf$io8$1@falcon.inetnebr.com> References: <341A2776.150A@afs.com> <5vds3m$di4$1@ironhorse.plsys.co.uk> <joe.ragosta-ya02408000R1309971445350001@news.dol.net> <5verp0$f7i$1@concorde.ctp.com> NNTP-Posting-User: bettis tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) writes: >Our software _must_ run on WinNT once it ships...but _none_ of the core >developers was willing to develop on NT... We tried that for a while, but it just didn't work. Just had to bite the bullet and go WinNT. Actually, we only have two intel Nextstep machines left. One is the firewall, and the other dual boots to Winnt, which is normally where it sits. >...so in order to test final look (some elemets change in size in the Win >theme...colors need to match etc) we simple run these apps with Win >Interface. >A really good reason if you ask me...you don't have to touch NT at all. Only >the guys we do system test have to run that thing. That wasn't enough for our software. There were too many subtil differences that could only be fixed by being there. Many things in the OpenStep classes are not completly defined. That is okay as long as you stick to what is documented and are not depending on undocument/unsupported behavior. When we switched to NT, all sorts of new bugs showed up. Little things that we had assumed worked a certain way, but didn't under NT. Like when you can draw and when a mouse click comes. For example, you can't remove a menu item from a menu while it is being clicked on under NT, but you can under Mach. >P.S. Want to hear some NT stories ? about swap files etc.. should we open a >new newsgroup for that. comp.sys.next.winstories :-) I think that is a good idea!
From: Chris Johnson <cjohnson@object-works.com> Newsgroups: comp.sys.next.programmer Subject: Re: Wait Cursor on NT Date: Mon, 15 Sep 1997 23:25:05 -0400 Organization: Erol's Internet Services Message-ID: <341DFC11.E1591F11@object-works.com> References: <5v8vc3$eb4$1@concorde.ctp.com> <5varvp$fjr@mochi.lava.net> <5vb5jk$7j6$1@concorde.ctp.com> <5veull$1fs$1@falcon.inetnebr.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sorry it took so long for me to get back to you: Basically this category should do what you need: #ifdef WIN32 #import <windows.h> //Apple secret NSCursor initializer @interface NSCursor (Secret) -(id)initWithCursorHandle:(HCURSOR)handle; @end #endif @implementation NSCursor (waitCursor) + (NSCursor *)waitCursor { static NSCursor *waitCursor = nil; #ifdef WIN32 if ( !waitCursor ) { HCURSOR cursor = LoadCursor(NULL, IDC_WAIT); if ( cursor && [NSCursor cursor instancesRespondToSelector:@selector(initWithCursorHandle:)] ) { waitCursor = [[NSCursor alloc] initWithCursor:cursor]; } } #endif return waitCursor; } @end In my NSApp delegate class I have an -(void)activateWaitCursor { [[NSCursor waitCursor] push]; } and -(void)deactivateWaitCursor { [[NSCursor waitCursor] pop]; } Call these methods before and after prints, saves, blah blah...Remember though that a modal panel seems to pop the waitCursor and that if you do display a modal panel (open, save, print, warning) after you push the waitCursor that you are going to send activateWaitCursor (sending push again) again after the modal session is over to put it back on the stack. needs only one deactivate since were sending the pop to the cursor vs. the cursor stack using the +pop method. Good luck. Chris Johnson Just wrote some init methods (no autorelease) for NSFileHandles today that may come in handy for those of you who are using the autoreleased convenience methods and are running out of filehandles because you have opened too many files (NT how grand) before the pool has released and returned the filehandles to the OS. Let me know if they are of use when and if you run into this problem. crj
From: Chris Johnson <cjohnson@object-works.com> Newsgroups: comp.sys.next.programmer Subject: EOF 2.1 Sybase 11 Adaptor problems Date: Mon, 15 Sep 1997 23:34:12 -0400 Organization: Erol's Internet Services Message-ID: <341DFE34.BBD6B5AF@object-works.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit CC: Robin Cohen <g8urzc@fanniemae.com> Just a heads up to let you know that the shipping Sybase EOAdaptor uses Sybase 10 client libraries. This works with Sybase 11.01 but when upgrading to 11.1 or above you get exceptions and the adaptor is broke. Word is that sybase changed from a Pascal stack convention to a C stack convention without telling anyone. Parameters for the connection dictionary get looked at in the wrong order. This has been fixed and there will be a patch available this week and I will try and get it stuck on NextAnswers.
From: Chris Johnson <cjohnson@object-works.com> Newsgroups: comp.sys.next.programmer Subject: Re: Wait Cursor on NT Date: Mon, 15 Sep 1997 23:40:03 -0400 Organization: Erol's Internet Services Message-ID: <341DFF93.7F6577DF@object-works.com> References: <5v8vc3$eb4$1@concorde.ctp.com> <5varvp$fjr@mochi.lava.net> <5vb5jk$7j6$1@concorde.ctp.com> <5veull$1fs$1@falcon.inetnebr.com> <341DFC11.E1591F11@object-works.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit typo in conditional statement...replace if ( cursor && [NSCursor cursor instancesRespondToSelector:@selector(initWithCursorHandle:)] ) with if ( cursor && [NSCursor instancesRespondToSelector:@selector(initWithCursorHandle:)] ) Chris Chris Johnson wrote: > Sorry it took so long for me to get back to you: > > Basically this category should do what you need: > > #ifdef WIN32 > #import <windows.h> > > //Apple secret NSCursor initializer > @interface NSCursor (Secret) > > -(id)initWithCursorHandle:(HCURSOR)handle; > > @end > #endif > > @implementation NSCursor (waitCursor) > > + (NSCursor *)waitCursor > { > static NSCursor *waitCursor = nil; > > #ifdef WIN32 > > if ( !waitCursor ) > { > HCURSOR cursor = LoadCursor(NULL, IDC_WAIT); > > if ( cursor && [NSCursor cursor > instancesRespondToSelector:@selector(initWithCursorHandle:)] ) > { > waitCursor = [[NSCursor alloc] initWithCursor:cursor]; > } > } > > #endif > > return waitCursor; > } > > @end > > In my NSApp delegate class I have an > > -(void)activateWaitCursor > { > [[NSCursor waitCursor] push]; > } > > and > > -(void)deactivateWaitCursor > { > [[NSCursor waitCursor] pop]; > } > > Call these methods before and after prints, saves, blah blah...Remember though > that a modal panel seems to pop the waitCursor and that if you do display a > modal panel (open, save, print, warning) after you push the waitCursor that you > are going to send activateWaitCursor (sending push again) again after the modal > session is over to put it back on the stack. needs only one deactivate since > were sending the pop to the cursor vs. the cursor stack using the +pop method. > > Good luck. > > Chris Johnson > > Just wrote some init methods (no autorelease) for NSFileHandles today that may > come in handy for those of you who are using the autoreleased convenience > methods and are running out of filehandles because you have opened too many > files (NT how grand) before the pool has released and returned the filehandles > to the OS. Let me know if they are of use when and if you run into this > problem. > > crj
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: How did Semper.fi die? (was Re: Apple developers continue to shift Followup-To: comp.sys.mac.advocacy Date: 16 Sep 1997 04:03:27 GMT Organization: none Message-ID: <5vl0ef$cku$5@ha2.rdc1.nj.home.com> References: <199709141729391030613@accs-as31-dp14.dlls.grid.net> <B041E51A-926E6@206.165.44.29> <SCOTT.97Sep15024112@howard.doubleu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NOTE: FOLLOWUPS TO COMP.SYS.MAC.ADVOCACY Remind me again why this thread is in comp.sys.next.programmer? Lawson ought to make everyone's killfile, or at least if you are going to listen to his rabid ravings, trim the followups..... Of course most Mac users I email about mis-posting tell me they had no idea where they were crossposting to.... TjL
From: Matt Watson <mgw@pacbell.net> Newsgroups: comp.sys.next.programmer Subject: Re: can't COMPILE DYNAMIC LIBRARIES Date: Mon, 15 Sep 1997 21:49:04 -0700 Organization: Thaumaturge, Inc. Message-ID: <341E0FC0.82@pacbell.net> References: <5up86i$pob$1@lyra.csx.cam.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Thomas Harte wrote: > HERE ARE THE PROBLEM LINES: > ============================ > ezee.h: > /* sizeof the image data header's records important for byte reversal */ > > size_t idcodefield = sizeof(int32); > size_t dsexfield = sizeof(int16); > size_t dtypefield = sizeof(EzeeType); > size_t dcomplexfield = sizeof(int16); > size_t nrowsfield = sizeof(int16); > size_t ncolsfield = sizeof(int16); > size_t nbandsfield = sizeof(int16); > size_t nsetsfield = sizeof(int16); > size_t data_sizefield = sizeof(ulong); You need to move these definitions into a .c file and change your .h file to declare them like: extern size_t idcodefield; extern size_t dsexfield; etc. Since you're including the .h file in multiple .c files, you get multiple common definitions of the same symbol. This is specified as an error by ANSI C. matt.
From: Rene Berber <_r.berber@computer.org_> Newsgroups: comp.sys.next.programmer Subject: Re: Memory leak when using Enterprise Objects Framework 1.1 Date: 16 Sep 1997 08:37:02 GMT Organization: None Whatsoever Message-ID: <5vlgfe$nt6$1@hp.fciencias.unam.mx> References: <5vin9d$j0s$1@news.NL.net> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: emmerik@hpb.holec-projects.nl P.J.L.van Emmerik wrote: > > I am using Enterprise Object Framework 1.1 to access an Oracle database. > There are some tables that, when displayed on screen, i always want to display the > 'current' state of the database. > I have tryd to sove this by doin a fetch from a timed entry. > When i do this i get a memory leak i cannot find. > I have tryd to use my own autorelease pool in the method that handles the timed entry. > I hav alse tryd the 2082_EOApplicationPatch.o patch. > Both didn't help. > > Any suggestions? > Sounds like you forgot to make the app an instance of EOApplication. In your app's main file change the class Application to EOApplication so you end with something resembling this: /* Generated by the NeXT Project Builder NOTE: Modified by hand so disable automatic main generation in PB! */ #import <appkit/appkit.h> #import <eointerface/EOApplication.h> void main(int argc, char *argv[]) { [EOApplication new]; if ([NXApp loadNibSection: "MAIN.nib" owner: NXApp withNames:NO]) [NXApp run]; [NXApp free]; exit(0); } This is needed to start the garbage collection used by EOF and your objects, otherwise all those autorelease messages don't really get rid of the objects and your app's memory use grows all the time. Hope this solves the problem, --- Rene Berber r.berber@computer.org (MIME/NeXTMail welcomed)
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C?... categories contra OOP ?? Date: 16 Sep 1997 10:48:14 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7vi01tzjl.fsf@tom.ics.ele.tue.nl> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> <3414041A.E71BFBCD@online.disney.com> <5vhgbr$c92$1@concorde.ctp.com> In-reply-to: tsengel@immd9.informatik.uni-erlangen.de's message of 14 Sep 1997 20:10:35 GMT In article <5vhgbr$c92$1@concorde.ctp.com> tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) writes: P.S. Btw.. what I would like to play around with are "context-sensitive" methods. Use C++, without using virtual methods. Play with it and implement a large project. Then come back and admit what you wished for was not what you wanted. You do not want any notion that the message sender has about your type to influence how the message is dispatched. --Tiggr
From: fairon@fltr.ucl.ac.be Newsgroups: comp.sys.next.programmer Subject: Problem with sed editor Date: Tue, 16 Sep 1997 11:16:59 +0000 Organization: Universite Catholique de Louvain, Louvain-la-Neuve, Belgium Message-ID: <341E6AA8.D76@fltr.ucl.ac.be> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hello, I try to edit a ASCII file with sed (i want to erase spaces between caracter "n" and caracter ":" in the ASCII file. Eg : "textn :") So I try to use the sream editor "sed", but it seems that the free space on the shell window dont't match the free space in the ASCII file. I tried sed 's/n ://g' file > newfile sed 's/n *://g' file > newfile ... but it did not match anything. Can s.o. help me to fix this problem ? Is it a problem of configuration of the shell windows ? Cédrick Fairon UCL/FLTR fairon@fltr.ucl.ac.be
From: Emmanuel Sciara <sciara@gatwick.geco-prakla.slb.com> Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: Re: prelude to Rhapsody on Sparc?!?! Date: Tue, 16 Sep 1997 11:12:16 +0000 Organization: Schlumberger Geco-Prakla Message-ID: <341E6990.446B9B3D@gatwick.geco-prakla.slb.com> References: <341E6923.2781E494@gatwick.geco-prakla.slb.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi all, Have I correctely understood??? Prelude to Rhapsody is available on Sun Sparc!!!??? How comes? Will Rhapsody be available on Sparc? How can I get Prelude to Rhapsody? E. -- +-----------------------------------------------------------------------------+ Emmanuel SCIARA Schlumberger Geco-Prakla Tel: (44)-(0)1293-55-6827 DAPD Field Processing Fax: (44)-(0)1293-55-6066 Schlumberger House Email: sciara@gatwick.Geco-Prakla.slb.com Gatwick Airport West Sussex RH6 0NZ, UK +-----------------------------------------------------------------------------+
From: "nala" <tlk684@modgnik.com> Newsgroups: comp.sys.next.hardware,comp.sys.next.marketplace,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.northstar,comp.sys.novell,comp.sys.nsc,comp.sys.nsc.32k,comp.sys.oric,comp.sys.os2,comp.sys.pal Subject: RE: Listen and you will smile... Date: 16 Sep 1997 07:40:22 GMT Organization: All USENET -- http://www.Supernews.com Message-ID: <01bcbec3$15d7dde0$b9bf39cc@default> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Lonely? Need to hear a soft, smiling voice? 1-900-255-0900, Ext 4096 $3.99 per min. Must be 18y. Serv-U (619) 645-8434 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Guillaume Proux <Guillaume.Proux@aar.alcatel-alsthom.fr> Newsgroups: comp.sys.next.programmer Subject: Re: Problem with sed editor Date: Tue, 16 Sep 1997 14:27:23 +0200 Organization: Alcatel Alsthom Recherche, Marcoussis, France Message-ID: <341E7B2B.70F4@aar.alcatel-alsthom.fr> References: <341E6AA8.D76@fltr.ucl.ac.be> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit fairon@fltr.ucl.ac.be wrote: > sed 's/n ://g' file > newfile > > sed 's/n *://g' file > newfile What is always use is: sed 's/\(.*\)n\(\ *\):/\1n:/g' I think it should work seamlessly! Guillaume
From: guyt@is.twi.tudelft.nl (A. Guyt) Newsgroups: comp.sys.next.programmer Subject: Rhapsody API's uploaded at peak Date: 16 Sep 1997 13:32:56 GMT Organization: Delft University of Technology Message-ID: <5vm1q8$a2$1@news.tudelft.nl> L.S., The two files containing the new Rhapsody API's for Foundation Kit and the Application Kit are available at the peak ftp site !! Address: ftp://next-ftp.peak.org/pub/openstep/rhapsody/sourcelibrary/docs/ files: - RhapsodyFoundationKit.tar.gz (1.7 MB) - RhapsodyApplicationKit.tar.gz (3.1 MB) Have e good time using them, Abraham Guyt. _____________________________________________________________________ Abraham Guyt P.O.Box 356 Department of Information Systems 2600 AJ Delft Faculty Information Technology & Systems The Netherlands Delft University of Technology tel: +31 15 278 5969 E-mail: guyt@is.twi.tudelft.nl NeXT-mail welcome
From: "Lawson English" <english@primenet.com> Newsgroups: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: How did Semper.fi die? (was Re: Apple developers continue to shift Date: 16 Sep 1997 08:00:01 -0700 Organization: Primenet Services for the Internet Message-ID: <B043F247-2F38C@206.165.44.44> References: <5vl0ef$cku$5@ha2.rdc1.nj.home.com> nntp://news.primenet.com/comp.sys.mac.programmer.misc, nntp://news.primenet.com/comp.sys.next.advocacy, nntp://news.primenet.com/comp.sys.next.programmer MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Timothy J. Luoma <nospam@all.please> said: > > Remind me again why this thread is in comp.sys.next.programmer? > Because NeXT developers are now Apple developers and deserve to have a say in commenting-on/learning-about why the quasi-official Apple developer feedback list died? ------------------------------------------------------------------------- "Apple's the Princess Diana of the computer industry." -Larry Campbell "And Steve Jobs has taken over for the regular driver." -Takechan <take@oatmeal.com> -------------------------------------------------------------------------
From: Emmanuel Sciara <sciara@gatwick.geco-prakla.slb.com> Newsgroups: comp.lang.c++,comp.sys.next.programmer Subject: C++ RTTI vs Objective-C Date: Tue, 16 Sep 1997 15:17:07 +0000 Organization: Schlumberger Geco-Prakla Message-ID: <341EA2F3.3F54BC7E@gatwick.geco-prakla.slb.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit If I am not mistaken, RTTI gives dynamic binding to C++. How is it compared to objective-C's own dynamic binding? Is it any difficult to use? Cheers E -- +-----------------------------------------------------------------------------+ Emmanuel SCIARA Schlumberger Geco-Prakla Tel: (44)-(0)1293-55-6827 DAPD Field Processing Fax: (44)-(0)1293-55-6066 Schlumberger House Email: sciara@gatwick.Geco-Prakla.slb.com Gatwick Airport West Sussex RH6 0NZ, UK +-----------------------------------------------------------------------------+
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.lang.c++,comp.sys.next.programmer Subject: Re: C++ RTTI vs Objective-C Date: 16 Sep 1997 17:35:58 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7wwkhi84h.fsf@tom.ics.ele.tue.nl> References: <341EA2F3.3F54BC7E@gatwick.geco-prakla.slb.com> In-reply-to: Emmanuel Sciara's message of Tue, 16 Sep 1997 15:17:07 +0000 In article <341EA2F3.3F54BC7E@gatwick.geco-prakla.slb.com> Emmanuel Sciara <sciara@gatwick.geco-prakla.slb.com> writes: If I am not mistaken, RTTI gives dynamic binding to C++. You are mistaken: every virtual member function is dynamically bound. You don't need RTTI for that. --Tiggr
From: Jim Redman <Jim@ergotech.com> Newsgroups: comp.sys.next.programmer Subject: DO Events and Modal Panels Date: Tue, 16 Sep 1997 11:09:44 -0600 Organization: ErgoTech Systems, Inc. Message-ID: <341EBD57.44A9@ergotech.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit CC: jim@ergotech.com In 3.3 the -runFromAppkitWithPriority methods for NXConnections would allow messages from DO to arrive and therefore the screen to be updated while modal panels were presented within the application or the mouse was down/window moving etc. etc.. Does anyone know a technique to achieve a similar effect under OpenStep? The default behavior whereby no DO messages arrive and then a whole bunch arrive sometime later seems to me to be completely crazy. Jim
From: tfu@bigfoot.com (Thomas F. Unke) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: Mon, 15 Sep 1997 21:29:41 GMT Organization: Smoke'n Joy - The first virtual crematory Message-ID: <1997Sep15.212941.1550@online.de> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: comrade@umich.edu In <B042DB4F-78B2@141.214.128.36> "Robert A. Decker" wrote: > You forgot one of the most important things: Not having to worry about > type in Obj-C. After programming in Obj-C and Newtonscript and a little > Dylan I absolutely cannot stand the static typing of Java/C++. > In fact, static typing ruins the concept of OO programming. Why should I > care what type the object is that I just got? Bertrand Meyer wrote about this subject (originally he talked about Smalltalk): "What about an airplane control system where suddenly a debugger-window pops up 'Message not understood' ?" With static typing the compiler ensures that all messages to an object can be handled. With Smalltalk/ObjC this is not the case. While static typing is often more clumsy than dynamic typing, it has its advantages.
From: markfr@galway.cse.tek.com (Mark S Frank) Newsgroups: comp.sys.next.programmer Subject: Help using <iostream.h> in OS NT Date: 16 Sep 1997 17:12:23 GMT Organization: Tektronix, Inc. Distribution: USA Message-ID: <5vmeln$gmm$1@bvadm.bv.tek.com> I am having problems including/using stream stuff such as #include <iostream.h> #include <fstream.h> on Openstep for NT. When I include these files, I get a compiler error when Microsoft's <ios.h> gets included in. I tried adding -lstdc++ to the makefile.preamble, but this didn't seem to help either. Anyone have any ideas? Also, how do I get a Complex C++ class?: - Is libg++ available on OS NT - I have tried including MS' <complex>, but the compiler can't find it. - Thanks for any help, Mark mark.s.frank@tek.com
From: markfr@galway.cse.tek.com (Mark S Frank) Newsgroups: comp.sys.next.programmer Subject: Links on NT Date: 16 Sep 1997 17:16:00 GMT Organization: Tektronix, Inc. Distribution: USA Message-ID: <5vmesg$gmm$2@bvadm.bv.tek.com> I would like to use the same set of C++ classes in a number of applications on Openstep for NT. I was able to do this using links on NS 3.3, but I have been unable to find links in NT. When I just try including these files into my project, Project Builder appears to copy the files into the directory. Any suggestions? - Thanks, Mark mark.s.frank@tek.com
Newsgroups: comp.sys.next.programmer From: mvlems@vbox.xs4all.nl (Mark F. Vlems) Subject: Problem with NSMutableString appendString in 4.0 Content-Type: text/plain; charset=us-ascii Message-ID: <1997Sep16.171958.16094@vbox.xs4all.nl> Sender: mvlems@vbox.xs4all.nl (Mark F. Vlems) Content-Transfer-Encoding: 7bit Mime-Version: 1.0 Date: Tue, 16 Sep 1997 17:19:58 GMT When I try to append a NSString to a NSMutableString I get an error. Maybe I am overlooking something, could anybody give me a hint? Thanks, Mark. Given a: -------- NSMutableString *value; -------- And a method: --------- - (id) appendValue :(const char *)valueToAppend { NSString *tmp; tmp = [[NSString alloc] initWithCString:valueToAppend]; [value appendString:tmp]; ... } --------- Gives me runtime the message: Sep 14 02:06:01 Test[8259] *** Uncaught exception: *** -[NSInlineCString appendString:]: selector not recognized exiting! -- Mark F. Vlems Poor is the man, MIME Mail OK Whose pleasures depend NeXTMail preferred! On the permission of another. Fax: +31 (0) 23 5622368 Madonna in "Justify my love", 1990
From: scott@doubleu.com (Scott Hess) Newsgroups: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: How did Semper.fi die? (was Re: Apple developers continue to shift Date: 16 Sep 1997 18:05:57 GMT Organization: Is a sign of weakness Message-ID: <SCOTT.97Sep16132146@slave.doubleu.com> References: <5vl0ef$cku$5@ha2.rdc1.nj.home.com> <B043F247-2F38C@206.165.44.44> In-reply-to: "Lawson English"'s message of 16 Sep 1997 08:00:01 -0700 In article <B043F247-2F38C@206.165.44.44>, "Lawson English" <english@primenet.com> writes: Timothy J. Luoma <nospam@all.please> said: > Remind me again why this thread is in comp.sys.next.programmer? Because NeXT developers are now Apple developers and deserve to have a say in commenting-on/learning-about why the quasi-official Apple developer feedback list died? Plus, NeXT developers are obviously too stupid to find the Apple developer newsgroups all by themselves, so we have to bring enlightenment to them forcibly. "have a say in commenting-on" indeed, -- scott hess <scott@doubleu.com> (606) 578-0412 http://www.doubleu.com/ <Favorite unused computer book title: The Demystified Idiots Guide to the Zen of Dummies in a Nutshell in Seven Days, Unleashed>
From: jmeacham@charm.net Newsgroups: comp.sys.next.programmer Subject: Anyone gotten Expect to compile? Date: 16 Sep 1997 20:29:51 GMT Organization: Charm.Net Baltimore Internet Access, Hon (410) 558-3900 Message-ID: <5vmq7v$7bs$1@news2.charm.net> Hi all, Being more of a sysadmin than a programmer, I'm having a hell of time getting Expect to compile on my ColorTurboStation with NS 3.3. HAs anyone gotten this to compile, and either want to tell me how or send me the binary via NeXTMail? I'd be eternally greatful. Peace, James -- ------------------------------------------------------------------- The Rev. James David Meacham Minister, Webmaster, Philosopher "If people did not sometimes do silly things, nothing intelligent would ever get done." -- Ludwig Wittgenstein
From: Raul Alvarez <ralvarez@sqi.com> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: Tue, 16 Sep 1997 14:39:06 -0700 Organization: Siemens Medical Systems, Inc. - Ultrasound Group Message-ID: <341EFC7A.167EB0E7@sqi.com> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Thomas F. Unke wrote: > "What about an airplane control system where suddenly a debugger-window pops > up 'Message not understood' ?" In Objective-C you can redefine this behavoir. I think it is called a 'PoseAs' object. > > With static typing the compiler ensures that all messages to an object can be > handled. With Smalltalk/ObjC this is not the case. While static typing is > often more clumsy than dynamic typing, it has its advantages. I think the proper nomencleture is 'run-time' vs. 'compile-time' binding. In Objective-C the programmer has the option to use 'compile-time' binding. If you specify the type as an (NSNumber *) you will not be able to compile if your methods arn't of an NSNumber or its superclasses. Raul.
Newsgroups: comp.lang.c++,comp.sys.next.programmer From: ray@netcom.com (Ray Fischer) Subject: Re: C++ RTTI vs Objective-C Message-ID: <rayEGMIEM.6nA@netcom.com> Organization: NETCOM, San Jose References: <341EA2F3.3F54BC7E@gatwick.geco-prakla.slb.com> Date: Tue, 16 Sep 1997 22:56:46 GMT Sender: ray@netcom4.netcom.com Emmanuel Sciara <sciara@gatwick.geco-prakla.slb.com> wrote: >If I am not mistaken, RTTI gives dynamic binding to C++. It does nothing of the sort. RTTI provides information about the polymorphic object, but does not affect the binding at all. In Objective-C, the method called for a particular object can be changed at runtime. >How is it compared to objective-C's own dynamic binding? >Is it any difficult to use? RTTI is a little obscure in its usage, but is not particularly difficult. Neither is it as useful as one might hope. -- Ray Fischer ray@netcom.com
From: spl@no.mo.spam.com Newsgroups: comp.sys.next.programmer Subject: OpenStep/Enterprise 4.1 linking problems Date: 16 Sep 1997 23:52:55 GMT Organization: BLaCKSMITH, Inc. Message-ID: <5vn64n$enf$1@anvil.BLaCKSMITH.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, I swear I've searched high and low for the answer to this simple, most basic of questions, but I have been unable to solve the problem. I have a single dll and it's header file and import library. I have a single source file (written in ansi C) that calls functions exported in the dll. The problem I have is that upon linking I get an error along the lines of : 'entrypoint mainCRT is not __stdcall with 12bytes of input' followed by a bunch of unresolved link errors (everyone of which should be defined in the import file). Setup: I created a new 'tool' project (the application is a command line program). removed the predefined main and replaced it with my existing c source code. Added the appropriate header directory entries in the Project inspector. Added the appropriate -L and -l entries in my Makefile.preamble (adding the through the 'libraries' interface resulted in gcc looking for <dllname>.lib.lib). No warnings when compiled with -Wall. just the silly link errors. Where did I go wrong? I can't believe I can't get something so simple to work properly. (forcing the creation of the exe by suppressing link errors (-Xlinker force or some such) results in a 'this is not a valid Windows NT application' error). Thanks much, -- I don't read the mail sent to the From: address in this post. If you would like to get a hold of me by email here's what you do: EITHER point your PH client to ns.uiuc.edu and query ian cardenas OR telnet to ux1.cso.uiuc.edu; login as phones; query ian cardenas
From: kennel@nospam.lyapunov.ucsd.edu (Matt Kennel (Remove 'NOSPAM' to reply)) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 17 Sep 1997 01:52:56 GMT Organization: University of California at San Diego Message-ID: <slrn61udth.6lt.kennel@lyapunov.ucsd.edu> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> <3414041A.E71BFBCD@online.disney.com> <5vh5bh$79q$1@walter.cs.umd.edu> On 14 Sep 1997 17:02:41 GMT, Sean Luke <seanlNoSpam@carmi.cs.umd.edu> wrote: :Joseph Panico (jpanico@online.disney.com) wrote: :>Here is what ObjC can do that Java cannot: :> :>2) Forwarding-- Forwarding is usefull for things like proxy classes and :>seemless method invocation failover. It's an elegant design pattern that :>can go a long way with a little creativity. Having it can conceptually :>simplify designs. : :While I'm also a big fan of Java, this point CANNOT be stressed heavily :enough. Java does not have sufficient guts to do distributed objects in :anything approaching the elegant way that Objective-C does (course, :garbage collection in distributed objects would also be a mess, but that's :another tale). In my opinion, this is a serious failing for a language :that claims to be truly dynamic. Sun's RMI tries very hard to provide rather decent distributed objects, including distributed garbage collection, 'stolen' from the Modula-3 research project. (Interesting that Sun is making $$$ out of the research paid for by DEC. This is why companies don't pay that much for basic research.) Obj-C does have lots of dynamic coolness, but Java was also designed to eventually be compilable 'down to the metal', which in truth means lots of compiletime cascading inlining and constant and type propagation. Substantial programmable redirection can be a hindrance to this. I don't think that Java is that great of a langauge either, they ignored lots of other good language research. :Sean Luke :seanl@cs.umd.edu -- * Matthew B. Kennel/Institute for Nonlinear Science, UCSD * * According to California Assembly Bill 3320, it is now a criminal offense * to solicit any goods or services by email to a CA resident without * providing the business's legal name and complete street address. *
From: Steve Dekorte <dekorte@slip.net> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 17 Sep 1997 02:02:07 GMT Organization: Slip.Net (http://www.slip.net) Message-ID: <5vndmv$2bl$1@owl.slip.net> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> Thomas F. Unke <tfu@bigfoot.com> wrote: > In <B042DB4F-78B2@141.214.128.36> "Robert A. Decker" wrote: > Bertrand Meyer wrote about this subject (originally he talked about > Smalltalk): > "What about an airplane control system where suddenly a debugger-window pops > up 'Message not understood' ?" Smalltalk and Objective-C can catch all such situations when they happen without "crashing"(so to speak) anything. Unlike Java/C++ there is a forward method in the root class. You can also check whether a target responds to method before sending it. > With static typing the compiler ensures that all messages to an object can be > handled. With Smalltalk/ObjC this is not the case. While static typing is > often more clumsy than dynamic typing, it has its advantages. Sure, but these are things that the dev environment should look for - not the compiler. The problem with static typing is that it makes too many assumptions. Really bad assumptions as it turns out. Like nothing will change at runtime. Like all target and actions are known at compile time. Like all actions and targets are in the processes address space. Etc, etc. I would bet that if you look at any large project implemented in both types of languages you'd find that the static side spends more time debugging all the extra code they need than the weak typing guys do debugging bad messages. -- Steve Dekorte - OpenStep consultant - San Francisco
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.next.programmer Subject: Re: Anyone gotten Expect to compile? Date: 17 Sep 1997 03:20:46 GMT Organization: none Message-ID: <5vniae$8ht$4@ha2.rdc1.nj.home.com> References: <5vmq7v$7bs$1@news2.charm.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: jmeacham@charm.net In <5vmq7v$7bs$1@news2.charm.net> jmeacham@charm.net wrote: > Hi all, > > Being more of a sysadmin than a programmer, I'm having a hell of time > getting Expect to compile on my ColorTurboStation with NS 3.3. HAs > anyone gotten this to compile, and either want to tell me how or send > me the binary via NeXTMail? I'd be eternally greatful. Peace, I've got the binary, but you need TCL 7.4 installed for it to run at all. If you have that, I can send you the binary... I'm not sure what version it is thou... I'd like to know if anyone else has gotten it to compile, I need an Intel version and it failed to compile (I think TCL even failed). TjL -- I'm sick of SPAM. Therefore, my email address is invalid. To email me, use one of these without spaces: luomat + next @ luomat.peak.org (for NeXT related stuff) luomat + peak @ luomat.peak.org (for PEAK related stuff)
Newsgroups: comp.sys.next.programmer From: mvlems@vbox.xs4all.nl (Mark F. Vlems) Subject: Re: Problem with NSMutableString appendString in 4.0 solved Content-Type: text/plain; charset=us-ascii Message-ID: <1997Sep17.052242.17325@vbox.xs4all.nl> Sender: mvlems@vbox.xs4all.nl (Mark F. Vlems) Content-Transfer-Encoding: 7bit Cc: mvlems@vbox.xs4all.nl References: <1997Sep16.171958.16094@vbox.xs4all.nl> Mime-Version: 1.0 Date: Wed, 17 Sep 1997 05:22:42 GMT In <1997Sep16.171958.16094@vbox.xs4all.nl> Mark F. Vlems wrote: > When I try to append a NSString to a NSMutableString I get an error. > Maybe I am overlooking something, could anybody give me a hint? > > Thanks, > Mark. > > Given a: > -------- > NSMutableString *value; > ... I had forgotten to change the initialization line: value = [[NSString alloc] initWithCString:... into: value = [[NSMutableString alloc] initWithCString:... after introducing the new method. I guess I am working too hard. -- Mark F. Vlems Poor is the man, MIME Mail OK Whose pleasures depend NeXTMail preferred! On the permission of another. Fax: +31 (0) 23 5622368 Madonna in "Justify my love", 1990
From: marcel@system.de Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 17 Sep 1997 06:06:08 GMT Organization: Technical University Berlin, Germany Distribution: world Message-ID: <5vns0g$27o$1@brachio.zrz.TU-Berlin.DE> References: <1997Sep15.212941.1550@online.de> In article <1997Sep15.212941.1550@online.de> tfu@bigfoot.com (Thomas F. Unke) writes: > In <B042DB4F-78B2@141.214.128.36> "Robert A. Decker" wrote: > > > You forgot one of the most important things: Not having to worry about > > type in Obj-C. After programming in Obj-C and Newtonscript and a little > > Dylan I absolutely cannot stand the static typing of Java/C++. > > In fact, static typing ruins the concept of OO programming. Why should I > > care what type the object is that I just got? > > Bertrand Meyer wrote about this subject (originally he talked about > Smalltalk): > > "What about an airplane control system where suddenly a debugger-window > pops up 'Message not understood' ?" The funny thing is that Eiffel's type system is not safe, so the same situation can occur. I don't know how Eiffel proceeds to handle such a situation, a core dump seems likely. Bertrand's response to the problems with his language's type system? 'Well, that never happens in practice'. Oops. > With static typing the compiler ensures that all messages to an > object can be handled. With Smalltalk/ObjC this is not the case. > While static typing is often more clumsy than dynamic typing, > it has its advantages. Actually, static typing is usually lumped together with 'class-based' typing, an unecessary restriction, whereas the 'message-based' typing that is more reasonable for object-oriented systems seems to be associated with dynamic typing, which is just as unecessary. It is perfectly possible to design type systems that are message-based, therefore allowing the full flexibility of OO-languages, but still checkable at compile time. In fact, Objective-C protocols offer most of this (except for after-the-fact supertypes) and both Smalltalk and C++ have been extended with these kinds of type systems. Marcel
From: marcel@system.de Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C?... categories contra OOP ?? Date: 17 Sep 1997 06:13:18 GMT Organization: Technical University Berlin, Germany Distribution: world Message-ID: <5vnsdu$2k0$1@brachio.zrz.TU-Berlin.DE> References: <x7vi01tzjl.fsf@tom.ics.ele.tue.nl> In article <x7vi01tzjl.fsf@tom.ics.ele.tue.nl> tiggr@ics.ele.tue.nl (Pieter Schoenmakers) writes: > > In article <5vhgbr$c92$1@concorde.ctp.com> tsengel@immd9.informatik.uni-erlangen.de (Thomas Engel) writes: > > P.S. Btw.. what I would like to play around with are "context-sensitive" > methods. > > Use C++, without using virtual methods. Play with it and implement a > large project. Then come back and admit what you wished for was not > what you wanted. What does that have to do with it? Context sensitivity as I understand it would provide more dynamic dispatch, not static. > You do not want any notion that the message sender has about your type > to influence how the message is dispatched. --Tiggr No I wouldn't. I would want to be able to specify the context in which this method is executed, with the method relying on its context to execute. I don't want this to be parameters because nesting should apply. Check out Northeastern's Adaptive Programming web site (context relation) or the work on nestable name-spaces etc. at the Vrije Universiteit Brussel. Marcel
From: pemmerik@solair1.inter.NL.net (P.J.L.van Emmerik) Newsgroups: comp.sys.next.programmer Subject: Re: Memory leak when using Enterprise Objects Framework 1.1 Date: Wed, 17 Sep 1997 06:32:33 GMT Organization: NLnet Message-ID: <5vnthk$18p$1@news.NL.net> References: <5vin9d$j0s$1@news.NL.net> <5vlgfe$nt6$1@hp.fciencias.unam.mx> Rene Berber <_r.berber@computer.org_> wrote: >P.J.L.van Emmerik wrote: >> >> I am using Enterprise Object Framework 1.1 to access an Oracle database. >> There are some tables that, when displayed on screen, i always want to >display the >> 'current' state of the database. >> I have tryd to sove this by doin a fetch from a timed entry. >> When i do this i get a memory leak i cannot find. >> I have tryd to use my own autorelease pool in the method that handles the >timed entry. >> I hav alse tryd the 2082_EOApplicationPatch.o patch. >> Both didn't help. >> >> Any suggestions? >> >Sounds like you forgot to make the app an instance of EOApplication. >In your app's main file change the class Application to EOApplication so you >end with something resembling this: >/* Generated by the NeXT Project Builder > NOTE: Modified by hand so disable automatic main generation in PB! >*/ >#import <appkit/appkit.h> >#import <eointerface/EOApplication.h> >void main(int argc, char *argv[]) >{ > [EOApplication new]; > if ([NXApp loadNibSection: "MAIN.nib" owner: NXApp withNames:NO]) > [NXApp run]; > > [NXApp free]; > exit(0); >} >This is needed to start the garbage collection used by EOF and your objects, >otherwise all those autorelease messages don't really get rid of the objects >and your app's memory use grows all the time. >Hope this solves the problem, >--- >Rene Berber >r.berber@computer.org >(MIME/NeXTMail welcomed) I use a my own EOApplication Class So in my case: void main(int argc, char *argv[]) { [MyApplication new]; if ([NXApp loadNibSection: "MAIN.nib" owner: NXApp withNames:NO]) [NXApp run]; [NXApp free]; exit(0); } @Interface MyApplication:EOApplication; ... @end So i assume this should work to. P.J.L. van Emmerik Holec Projects B.V. Email: emmerik@hpb.holec-projects.nl PO.BOX 565, 7550 AN Hengelo pemmerik@solair1.inter.NL.net The Netherlands Phone: +31 74 2558 688 --
Subject: Re: Is Java Better than Objective-C? Newsgroups: comp.sys.next.programmer References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> In-Reply-To: <1997Sep15.212941.1550@online.de> From: marco@sente.ch.mil (Marco Scheurer) Organization: Sen:te Message-ID: <341f8b03.0@epflnews.epfl.ch> Date: 17 Sep 97 07:47:15 GMT On 09/16/97, Thomas F. Unke wrote: > [...] >Bertrand Meyer wrote about this subject (originally he talked >about Smalltalk): > >"What about an airplane control system where suddenly a > debugger-window pops up 'Message not understood' ?" > What about an airplane control system where suddenly a debugger-window pops up 'Uncaught exception raised' ? This is just an example of the stupid things that Bertrand Meyer is capable of saying to prove that Eiffel is the only true object-oriented language, the only safe language, etc... Recently, he "proved" that if Eiffel had been used instead of Ada, the Ariane-5 crash (a $500 million software failure) would not have happened. As much as I admire some aspects of Eiffel and the work of Meyer, this is total BS. -- Marco Scheurer Sen:te (remove "dot mil" from my address)
From: gerriet@hazel.north.de (Gerriet M. Denkmann) Newsgroups: comp.sys.next.programmer Subject: Api for Edit.app Date: 17 Sep 1997 07:20:57 GMT Organization: Oldenburger Informations-Systeme, FRG Distribution: world Message-ID: <5vo0cp$1ci@voyager.north.de> I want some application of mine to interface with Edit.app. More specifically I want it to: 1. open a file in Edit.app (here I have found openFile:withApplication:) 2. send Edit some message (for this window just opened) equivalent to Edit -> Find -> Line Range... -> Line -> Select. In other words: I want to send some massage (to the Workspace Manager ?) like "tell Edit to select line nnn in the window with the file fff" The 3.3 ProjectBuilder can do it, so it can be done. But where is the Edit API defined? Or is there some source code available, which interfaces with Edit in this way, so that I can use it as an example? Gerriet.
From: szallies@energotec.de (Constantin Szallies) Newsgroups: comp.sys.next.programmer Subject: Re: Problem with NSMutableString appendString in 4.0 Date: 17 Sep 1997 07:06:16 GMT Organization: Technet GmbH Message-ID: <5vnvh8$i72$1@oxygen.technet.net> References: <1997Sep16.171958.16094@vbox.xs4all.nl> mvlems@vbox.xs4all.nl (Mark F. Vlems) wrote: >When I try to append a NSString to a NSMutableString I get an error. >Maybe I am overlooking something, could anybody give me a hint? > >Thanks, >Mark. > >Given a: >-------- > NSMutableString *value; >-------- > >And a method: >--------- >- (id) appendValue :(const char *)valueToAppend >{ > NSString *tmp; > > tmp = [[NSString alloc] initWithCString:valueToAppend]; > [value appendString:tmp]; > ... >} >--------- >Gives me runtime the message: > Sep 14 02:06:01 Test[8259] *** Uncaught exception: *** > -[NSInlineCString appendString:]: selector not recognized > exiting! The object referenced by value has the concrete class NSInlineCString which is a subclass of NSString but not NSMutableString. A NSInlineCString is immutable and was created by @"". You defined the type of the value reference as NSMutableString, but no type checking is done during the executed assignment of this reference. This compiles without warning: id aString = @"Hello"; NSMutableString *value = aString; The type is only checked at compile time. Greetings -- # Constantin Szallies, Energotec GmbH # szallies@energotec.de # http://www.energotec.de/~szallies/ # 49211-9144018
From: not@my.address.net Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 17 Sep 1997 15:09:02 GMT Organization: University of Michigan ITD News Server Message-ID: <5vorqe$5gb$1@newbabylon.rs.itd.umich.edu> References: <01bcb30b$32787ae0$0100000a@ivanova> <EFL8FL.7oE@basil.icce.rug.nl> <01bcb4cf$9b253c70$0100000a@ivanova> <3414041A.E71BFBCD@online.disney.com> In-Reply-To: <3414041A.E71BFBCD@online.disney.com> On 09/08/97, Joseph Panico wrote: > >Here is what Java can do that ObjC cannot: > >4) Objects only! aside from a small number of primitive classes, Java is >a pure OOP. No more legacy c stuff- structs pointers, etc. This vastly >increases the robustness of Java over ObjC. It also conceptually >simplifies the language. >5) Package(ing) of name space. Packages elegantly solve ObjC's >irritating problem of having a global Object name space. No more worries >of name collision and no more nonsensical naming conventions. >6) Dynamic class loading! This is a huge win! No more minute plus link >phases. This is why I dread programming ObjC. The combination of the >compiler and linker are about 100 times slower than the combination of >compile/load using Symantec Cafe. The current crop of Java environments >(Cafe, Asymmetrix, VisualAge) are so fast, that you are essentially >working in an interpreted environment. Re (4): If you're programming using the Foundation Kit and you're not working with legacy C code, can't you pretty much avoid C-stuff, ie structures, pointers, etc? That was my hope, at any rate ... Re (5): I've never understood how to think about the scope of object names in a language like Objective-C. Are they local to the methods that declare them in the same way that names of automatic variables are local to the functions and blocks that declare them in C? And should I think of class names the same way I would think about C structures having global scope? If that were right, there wouldn't be naming collisions between objects, only classes. But I'm not sure I really understand how to think about objects and classes as compared to C variables with local, file or global scope. If anyone could help me out on this point, I would really appreciate it. Re (6): Wasn't NeXT working on "Fix'n Go" as a solution to this problem? Earlier this year there a job posting at Apple mentioned the possibility of working on this technology. But it wasn't clear whether Apple was undecided about their commitment to the project itself or merely undecided whether the candidate for that particular position would be involved. Thanks to everyone that's contributed here. I'm curious about this stuff, but it's way out of my line of work. -- invert: umich.edu jdevlin insert: shift "2"
From: Alex Blakemore <alex@genoa.com> Newsgroups: comp.sys.next.programmer Subject: Re: Problem with NSMutableString appendString in 4.0 Date: 17 Sep 1997 02:42:20 GMT Organization: Genoa Software Systems Message-ID: <5vng2c$oa@saturn.genoa.com> References: <1997Sep16.171958.16094@vbox.xs4all.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: mvlems@vbox.xs4all.nl In <1997Sep16.171958.16094@vbox.xs4all.nl> Mark F. Vlems wrote: > When I try to append a NSString to a NSMutableString I get an error. > Maybe I am overlooking something, could anybody give me a hint? > NSMutableString *value; This line just declares to the compiler that value points to a NSMutableString, It doesn't by itself ensure anything at runtime (but it is good to declare types to get compiler help) > [value appendString:tmp]; > -[NSInlineCString appendString:]: selector not recognized This run time message means that what value really points to is an instance of NSInlineCString (which is a subclass of NSString, but not NSMutableString) So you need to investigate to make sure that you really set value to point to a MUTABLE string. -- Alex Blakemore alex@genoa.com NeXT, MIME and ASCII mail accepted
From: tom@basil.icce.rug.dev.null.nl (Tom Hageman) Newsgroups: comp.sys.next.programmer Subject: Re: Memory leak when using Enterprise Objects Framework 1.1 Date: 17 Sep 1997 17:41:46 GMT Organization: Warty Wolfs Sender: news@basil.icce.rug.nl (NEWS pusher) Message-ID: <EGMGIK.Hu3@basil.icce.rug.nl> References: <5vin9d$j0s$1@news.NL.net> pemmerik@solair1.inter.NL.net (P.J.L.van Emmerik) wrote: > > I am using Enterprise Object Framework 1.1 to access an Oracle database. > There are some tables that, when displayed on screen, i always want to display the > 'current' state of the database. > I have tryd to sove this by doin a fetch from a timed entry. > When i do this i get a memory leak i cannot find. > I have tryd to use my own autorelease pool in the method that handles the timed entry. > I hav alse tryd the 2082_EOApplicationPatch.o patch. > Both didn't help. > > Any suggestions? It seems that EOApplication accumulates stuff autoreleased in timed entries (or DO clients that -runFromAppKit for that matter) in its top-level autorelease pool until a user-event comes by. Needless to say this stinks:-p For my (non-EO, DO) application, I was able to hack around this by using the following category on PoolApplication (from NeXTanswer #1722), and then invoke [NXApp setPoolCleanupInterval:] with a suitable constant in -appWillInit:. I have no idea if it's feasible to do something similar on EOApplication, or that PoolApplication will work in EO apps... @interface PoolApplication (_releasePool) - (void)_releasePool; @end @implementation PoolApplication (TimedEntryPatch) /* Timed entry to regularly clean up the autorelease pool. This is for applications that handle stuff in timed entries, or are DO clients, that can otherwise accumulate a lot of junk in their top-level autorelease pool if no user-intervention occurs. */ static DPSTimedEntry tePool; static void tePoolHandler(DPSTimedEntry teNum, double now, void *self) { [(id)self _releasePool]; } - (void)setPoolCleanupInterval:(double)sec { if (tePool) DPSRemoveTimedEntry(tePool); if (sec > 0) tePool = DPSAddTimedEntry(sec, tePoolHandler, self, NX_BASETHRESHOLD); } @end // PoolApplication (TimedEntryPatch) Another workaround (that I haven't tried) may be to send a user-defined event after you've done your regular stuff in the timed entry, in the hopes that this would release the top-level autorelease pool. Hope this helps, Tom. [posted & mailed] -- __/__/__/__/ Tom Hageman <tom@basil.icce.rug.dev.null.nl> [NeXTmail/Mime OK] __/ __/_/ IC Group <tom@icgned.dev.null.nl> (work) __/__/__/ <<SPAMBLOCK: remove .dev.null to reply>> __/ _/_/ Confused? You won't be after the NeXT episode.
From: John Stanhope <john.stanhope@natinst.com> Newsgroups: comp.sys.next.programmer Subject: Re: Anyone gotten Expect to compile? Date: Wed, 17 Sep 1997 12:55:48 -0500 Organization: National Instruments Message-ID: <342019A4.F0AC1FE6@natinst.com> References: <5vmq7v$7bs$1@news2.charm.net> <5vniae$8ht$4@ha2.rdc1.nj.home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Timothy J. Luoma wrote: > In <5vmq7v$7bs$1@news2.charm.net> jmeacham@charm.net wrote: > > Hi all, > > > > Being more of a sysadmin than a programmer, I'm having a hell of time > > getting Expect to compile on my ColorTurboStation with NS 3.3. HAs > > anyone gotten this to compile, and either want to tell me how or send > > me the binary via NeXTMail? I'd be eternally greatful. Peace, > > I've got the binary, but you need TCL 7.4 installed for it to run at all. > > If you have that, I can send you the binary... I'm not sure what version it > is thou... > > I'd like to know if anyone else has gotten it to compile, I need an Intel > version and it failed to compile (I think TCL even failed). > To get any of the latest version of Tcl to compile 7.6 or 8.0 you need to add some macros to tclUnixFCmd.c from one of the ifdeffed out posix sections of <sys/file.h> or something (I am not in front of my NeXT) and remove the "-lc" flag from the make file generated after its configured. After that it will compile and run although I haven't tested it yet much. If there is interest I could put up some binaries on peak or peanuts. John Stanhope
From: tfu@bigfoot.com (Thomas F. Unke) Newsgroups: comp.sys.next.programmer Subject: Re: Api for Edit.app Date: Wed, 17 Sep 1997 20:24:05 GMT Organization: Smoke'n Joy - The first virtual crematory Message-ID: <1997Sep17.202405.1004@online.de> References: <5vo0cp$1ci@voyager.north.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: gerriet@hazel.north.de In <5vo0cp$1ci@voyager.north.de> Gerriet M. Denkmann wrote: > The 3.3 ProjectBuilder can do it, so it can be done. > > But where is the Edit API defined? > Yes, it can be done. I used it some years ago with the Listener/Speaker classes. here is a > forward method in the root class. You can also check whether a target > responds to method before sending it. > > > With static typing the compiler ensures that all messages to an object can be > > handled. With Smalltalk/ObjC this is not the case. While static typing is > > often more clumsy than dynamic typing, it has its advantages. > > Sure, but these are things that the dev environment should look for - > not the compiler. How does a dev environment handle "Message not understood" ? Having used NeXT-Apps for some years, one of the major reasons for a crash of an App was exactly this. Of course you can handle this, but the average (all) programmer does not always think of the possibility of an illegal message. In case of a dynamically typed language like ObjC or Smalltalk, to be sure that you have not overseen an illegal message, you have to test/debug all cases of program flow. This is possible to some extent for small or medium sized applications. But huge applications like an airplane control system, are just to big to handle this kind of tests.
From: nospam@all.please (Timothy J. Luoma) Newsgroups: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: How did Semper.fi die? (was Re: Apple developers continue to Followup-To: comp.sys.mac.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.advocacy Date: 18 Sep 1997 02:20:23 GMT Organization: none Message-ID: <5vq357$3ah$1@ha2.rdc1.nj.home.com> References: <5vl0ef$cku$5@ha2.rdc1.nj.home.com> <B043F247-2F38C@206.165.44.44> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: english@primenet.com NOTE: FOLLOWUPS TO COMP.SYS.MAC.ADVOCACY,COMP.SYS.MAC.PROGRAMMER.MISC,COMP.SYS.NEXT.ADVOCACY In <B043F247-2F38C@206.165.44.44> "Lawson English" wrote: > Timothy J. Luoma <nospam@all.please> said: > > > > Remind me again why this thread is in comp.sys.next.programmer? > > > Because NeXT developers are now Apple developers and deserve to have a say > in commenting-on/learning-about why the quasi-official Apple developer > feedback list died? 1) Who said that NeXT developers are now Apple developers? Who said that they CARE WHAT YOU ARE SPEWING? 2) EVER HEARD OF A FOLLOWUP-TO LINE? You deliberately posted an offtopic post BACK to the group after I had SPECIFICALLY removed it. KEEP THIS OUT OF THE PROGRAMMER GROUP! ALL FURTHER POSTS FROM YOU WILL BE RESENT TO YOU AND YOUR POSTMASTER AS OFF-TOPIC AS PER THE CHARTER. comp.sys.next.programmer Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. The NEXTSTEP programmer FAQs are posted here. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. comp.sys.next.advocacy This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. You are breaking netiquette all over the place, so grow up, lighten up, and figure out where to post to where there is a chance, albeit minute, that someone might actually care. I already have the tools in place to automate this process if you knowingly continue to break the charter. Your ISP has an AUP which has provisions for dealing with off-topic posts. TjL -- luomat + peak@luomat.peak.org Anything relating to next-ftp.peak.org luomat + next@luomat.peak.org NeXTStep/OpenStep related
Newsgroups: comp.sys.next.programmer,comp.sys.next.misc From: dfevans@bbcr.uwaterloo.ca (David Evans) Subject: Re: prelude to Rhapsody on Sparc?!?! Sender: news@novice.uwaterloo.ca (Mr. News) Message-ID: <EGnt7v.2rL@novice.uwaterloo.ca> Date: Wed, 17 Sep 1997 15:47:55 GMT References: <341E6923.2781E494@gatwick.geco-prakla.slb.com> <341E6990.446B9B3D@gatwick.geco-prakla.slb.com> Organization: University of Waterloo In article <341E6990.446B9B3D@gatwick.geco-prakla.slb.com>, Emmanuel Sciara <sciara@gatwick.geco-prakla.slb.com> wrote: > >Hi all, > >Have I correctely understood??? Prelude to Rhapsody is available on Sun >Sparc!!!??? Apparently, yes. However, I've heard that the CD isn't bootable, so you have to use a 3.3 CD or some such. >How comes? Will Rhapsody be available on Sparc? Doubtful. -- David Evans (NeXTMail/MIME OK) dfevans@bbcr.uwaterloo.ca Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/ University of Waterloo "Default is the value selected by the composer Ontario, Canada overridden by your command." - Roland TR-707 Manual
From: Curtis Crowson <curtis_crowson@emory.org> Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: Re: prelude to Rhapsody on Sparc?!?! Date: Thu, 18 Sep 1997 09:01:48 -0400 Organization: Emory University System of Health Care Message-ID: <3421263C.148F@emory.org> References: <341E6923.2781E494@gatwick.geco-prakla.slb.com> <341E6990.446B9B3D@gatwick.geco-prakla.slb.com> <EGnt7v.2rL@novice.uwaterloo.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > >How comes? Will Rhapsody be available on Sparc? > > Doubtful. > > -- Check out www.macosrumors.com, they mention that some internal documents about Rhapsody mention bugs in drivers to other platforms like Sun, and HP. Well here is the exact quote. Is Rhapsody quietly supporting HP-UX and Solaris? While sifting through stacks of documents relating to Rhapsody DR1, Rumors has come upon information that suggests that Apple developers may be working on versions of the Yellow Box APIs for Hewlett-Packard's HP-UX and Sun Microsystems' Solaris operating systems. There are a number of references to bugs found on the Yellow Box running on those systems, as well as planned fixes. This means that, at the very least, Apple is investing time and effort into investigating the possibilities of supporting those platforms -- a major boon for any developer planning applications for the Yellow Box. Here is the URL if you want to check it out yourself. http://www.macosrumors.com/archive57.shtml Of course being the rumor page you should not bet your software on it.
From: holger@object-factory.REMOVETHIS.com (Holger Hoffstaette) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 18 Sep 1997 10:00:44 GMT Organization: Object Factory GmbH (Germany) Message-ID: <5vqu4c$srg$1@leonie.object-factory.com> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Thomas F. Unke wrote: > In case of a dynamically typed language like ObjC or Smalltalk, to be sure > that you have not overseen an illegal message, you have to test/debug all > cases of program flow. This is possible to some extent for small or medium > sized applications. But huge applications like an airplane control system, > are just to big to handle this kind of tests. Yes, *in theory*. In practice, the 'message not understood' problem is far less serious than most people think. It really is. Please ask on comp.lang.smalltalk. The reasons are too numerous to be mentioned here, but involve, among other things, more 'natural' (i.e. based on an object's logic) message flow and thus radically decreased chances of wrong objects being messaged in the first place. Btw: a lot of airplane control systems are so huge not only because they are so inherently complex (they surely are!), but also because they have accumulated an amazing amount of historical cruft and often contain a great amount of duplicated functionality, e.g. for failover handling (separated subsystem control logic etc.). No matter what B. Meyer and the Static Typing Police try to tell you: there is very little reason to believe that a dynamically typed and/or bound system is *intrinsically* unsafer than something statically typed. I have no problem arguing that the possibility to dynamically re-route message flow is, in a way, more flexible and adaptable to rapid, unforeseen change than a system where every conceivable situation *must* be known in advance. The static typing of ObjC, together with the protocol-based dynamic capabilities, provide a very, very good compromise and let the programmer chose the right way for the job. That's the way it should be - reality is smarter than all of us anyway. All you can do is handle it as good as you can, and be prepared for the worst. Holger "Real programmers do it..dynamically." -- holger"at"object-factory.com (NeXTMail, MIME) 'Where I come from it's important to develop stuff that doesn't suck.' -- David Young <dwy"at"ace.net>
From: embuck@palmer.cca.rockwell.com (Erik M. Buck) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 18 Sep 1997 13:34:45 GMT Organization: Rockwell Collins Message-ID: <5vrall$ghn1@castor.cca.rockwell.com> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: tfu@bigfoot.com In <1997Sep17.201838.942@online.de> Thomas F. Unke wrote: > In <5vndmv$2bl$1@owl.slip.net> Steve Dekorte wrote: > > In case of a dynamically typed language like ObjC or Smalltalk, to be sure > that you have not overseen an illegal message, you have to test/debug all > cases of program flow. This is possible to some extent for small or medium > sized applications. But huge applications like an airplane control system, > are just to big to handle this kind of tests. > As a maker of avionics software, I assure you that absolutely every possible program flow is tested and when possible mathematically proven! The REAL reason that dynamic languages are not used in embedded avionics systems stems from the fact that for safety reasons, we do not allocate memory from the heap. Even the program stack is monitored (verified/proven) never to exceed design constraints. Avionics applications model well understood static relationships that NEVER change in the 50 year life of an airplane. Dynamic languages excel at all of the other problems that are by necessity messy, subject to frequent change, or not well understood in advance. For example, the Objective-C message dispatch could probably never be proven for use in airplanes because it uses (difficult to prove) hash tables and dynamic allocation. Different problems benefit from different languages. "Closed world" problems are best solved with static languages. "Open world" problems (the vast majority) such as GUIs, all of Openstep, practically every windows application etc. benefit from dynamic languages. >Having used >NeXT-Apps for some years, one of the major reasons for a crash of an App was >exactly this. Having used NeXT-Apps and non NeXT-Apps for 15 years, NeXT-Apps (even freeware) are MUCH more reliable than alternatives. The real question is "Why do so many C++ apps crash so frequently in spite of the fact that 'message not understood' is not one of the reasons ?"
From: ians@cam-ani.co.uk (Ian Stephenson) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: Thu, 18 Sep 1997 14:44:07 GMT Organization: Cambridge Animation Systems Ltd Sender: news@cam-ani.co.uk Message-ID: <EGpKxK.HEp@cam-ani.co.uk> References: <5vrall$ghn1@castor.cca.rockwell.com> In article <5vrall$ghn1@castor.cca.rockwell.com> embuck@palmer.cca.rockwell.com (Erik M. Buck) writes: > The real question is > "Why do so many C++ apps crash so frequently in spite of the fact that > 'message not understood' is not one of the reasons ?" I'd sugguest that one of the reasons is that 'message not understood' occurs MORE oftern in C++. It's just that static binding hides it so you get mysterious bad things happening rather than a clean diagnostic printed out. You take something out of a list in C++, and by static typing you know it's base type - oftern the equivalent of Object. You then have to hard code your assuptions as it's derived class. Dynamic binding will adapt to support any object which has the required messages, and hence will generally do the right thing. If I've had to explicitly cast it to a derived class (say because it's stored in a list in MY base class), then this is doomed to fail when the assumptions I've made become invalid. eg. my_base -> partner_base | | my_derived partner_derived my_derived member function calls inq_partner() to itself, and gets a pointer to a partner_base object. In practise I _KNOW_ that my_derived are always linked to partner_derived's so if I want to use the extended functionality I cast partner_base to partner_derived. Too bad when someone writes alt_partner_derived. Lack of introspection prevents things like w_assert([my_partner isKindof: partner_derived]; Worse than that, it might work for a while depending on how the pointers are layed out. If you did call a bad method you MIGHT get a crash, or you might not. Either way it's VERY hard to track down. It all comes back to the fact that Inheritance (an implementation detail) is not a valid mechanism for typing (an interface [somewhat more than ] detail). Type should be defined somewhat like protocols. Implementation should be totally hidden. $an
From: tom@basil.icce.rug.dev.null.nl (Tom Hageman) Newsgroups: comp.sys.next.programmer Subject: Re: Api for Edit.app Date: 18 Sep 1997 17:41:40 GMT Organization: Warty Wolfs Sender: news@basil.icce.rug.nl (NEWS pusher) Message-ID: <EGo419.L1D@basil.icce.rug.nl> References: <5vo0cp$1ci@voyager.north.de> gerriet@hazel.north.de (Gerriet M. Denkmann) wrote: > I want some application of mine to interface with Edit.app. > More specifically I want it to: > > 1. open a file in Edit.app (here I have found openFile:withApplication:) > > 2. send Edit some message (for this window just opened) equivalent to Edit > -> Find -> Line Range... -> Line -> Select. > In other words: I want to send some massage (to the Workspace Manager ?) > like "tell Edit to select line nnn in the window with the file fff" > > The 3.3 ProjectBuilder can do it, so it can be done. > > But where is the Edit API defined? > > Or is there some source code available, which interfaces with Edit in this > way, so that I can use it as an example? Well, there's always class-dump: % class-dump /NextApps/Edit.app/Edit [...] @interface EditListener:Listener [...] - (int)openFile:(STR)fp16 onHost:(STR)fp20 fromTrueLine:(int)fp24 to:(int)fp28; - (int)openFile:(STR)fp16 onHost:(STR)fp20 atTrueLine:(int)fp24; @end An example of its use can be found in Christian Limpach's open(1) replacement (simplified): char *fileName; int oline; char *hostName = NULL; Speaker *appSpeaker = [[Speaker alloc] init]; port_t appPort; // eventually set hostName. if((appPort=NXPortFromName("Edit", hostName)) != PORT_NULL) { [appSpeaker setSendPort:appPort]; [appSpeaker selectorRPC:"openFile:onHost:atTrueLine:" paramTypes:"cci", fileName, hostName?hostName:"localhost", oline]; } The Speaker and Listener classes are only available in NeXTSTEP, not in OPENSTEP. (Well, that's not entirely true. They are still declared in <AppKit/obsoleteSpeaker.h> and <AppKit/obsoleteSpeaker.h>, at least in OS/Mach.) The Speaker/Listener classes also promise the (documented) method: " - (int)msgSetPosition:(const char *)aString posType:(int)anInt andSelect:(int)selectFlag ok:(int *)flag Receives a remote message requesting the application to scroll the current document (the one displayed in the main window) so that the portion described by aString is visible. aString should be interpreted according to the anInt constant, which will be one of the following: NX_TEXTPOSTYPE aString is a character string to search for. NX_REGEXPRPOSTYPE aString is a regular expression to search for. NX_LINENUMPOSTYPE aString is a colon-separated range of line numbers, for example "10:12". NX_CHARNUMPOSTYPE aString is a colon-separated range of character positions, for example "21:33". NX_APPPOSTYPE aString is an application-specific description of a portion of the document. The msgSetPosition:posType:andSelect:ok: method you implement should set the integer referred to by flag to YES if the document is scrolled, and to NO if it isn't. If selectFlag is anything other than 0, the portion of the document described by aString should also be selected. " I haven't tried this method myself, so I don't know firsthand how well this works... Hope this helps, Tom. -- __/__/__/__/ Tom Hageman <tom@basil.icce.rug.dev.null.nl> [NeXTmail/Mime OK] __/ __/_/ IC Group <tom@icgned.dev.null.nl> (work) __/__/__/ <<SPAMBLOCK: remove .dev.null to reply>> __/ _/_/ Confused? You won't be after the NeXT episode.
From: pxpst2@vms.cis.pitt.edu (Peter) Newsgroups: comp.sys.next.programmer Subject: SMP and Rhapsody Date: Thu, 18 Sep 1997 14:34:14 -0400 Organization: U. of Pitt. Patholgy Message-ID: <pxpst2-1809971434140001@pelli.pathology.pitt.edu> How hard will the task of implementing SMP in Rhapsody be? Peter -- "Don't you eat that yellow snow watch out where the huskys go"
Message-ID: <3421B435.704@ibm.net> Date: Fri, 19 Sep 1997 00:07:33 +0100 From: Ralph Paul <repaul@ibm.net> MIME-Version: 1.0 Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> <5vqu4c$srg$1@leonie.object-factory.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Holger Hoffstaette wrote: > Yes, *in theory*. In practice, the 'message not understood' problem > is far less serious than most people think. It really is. > Please ask on comp.lang.smalltalk. The reasons are too numerous to > be mentioned here, but involve, among other things, more 'natural' > (i.e. based on an object's logic) message flow and thus radically > decreased chances of wrong objects being messaged in the first place. > Btw: a lot of airplane control systems are so huge not only because > they are so inherently complex (they surely are!), but also because > they have accumulated an amazing amount of historical cruft and often > contain a great amount of duplicated functionality, e.g. for failover > handling (separated subsystem control logic etc.). Sorry this sounds like you have not seen a real flight control or avionics system. While flight control system might be the result of many years of development, they certainly do not contain any code that unnecessary since runtime performance is cirtical and processor resources are always something that's too small. Also you have to keep in mind that avionics or flight control system have to be proven to work correctly not just by doing runtime tests but on paper. Therefore Obj-C or C++ or Smalltalk or Eiffel are rather unlikely options. Ada83 is commonly used due to it's rigid static typing. > No matter what B. Meyer and the Static Typing Police try to tell > you: there is very little reason to believe that a dynamically > typed and/or bound system is *intrinsically* unsafer than something > statically typed. I have no problem arguing that the possibility > to dynamically re-route message flow is, in a way, more flexible > and adaptable to rapid, unforeseen change than a system where every > conceivable situation *must* be known in advance. > The static typing of ObjC, together with the protocol-based dynamic > capabilities, provide a very, very good compromise and let the > programmer chose the right way for the job. That's the way it should > be - reality is smarter than all of us anyway. All you can do is > handle it as good as you can, and be prepared for the worst. Final comment : The flight control example from OOSC 2 ( dynamic <-> static typing ) is really a bad one because avionics, flight control systems are developed in a very different manner than normal applications. B. Meyer should have made his comments using a different example Ralph Paul
From: Trevin Beattie <*trevin*@*xmission*.*com*> Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE ------- BOARD!! Date: Thu, 18 Sep 1997 21:51:32 -0600 Organization: XMission Internet (801 539 0900) Message-ID: <5vsss4$3u2$1@news.xmission.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <don_arb-0509971214210001@sea-ts1-p59.wolfenet.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Don Arbow wrote: > > Most of the hype about NC's deals with corporations and large intranets > where employees are using $2500 PC workstations to play solitaire. The > main push behind using NC's is that these companies can get rid of those > pricy machines, if the users are only going to run a word processor and a > database client. Centralized storage for programs and data makes sense > for maintenance and administration costs. The company that I used to work for employed around a thousand technicians and customer support people, and we all depended on our computers heavily. One of our biggest complaint for years was the network being too slow; all our apps and shared data was stored on the network server. Finally, when management decided to budget the money for it, we had all our computers upgraded with larger hard drives and the applications were installed locally. It made quite a big difference. Shared data and important files were still kept on the server, but the network wasn't bogged down as much when people opened their daily programs. Just my 2¢.
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <5vl1k2$nn9$525@news.insnet.com> Control: cancel <5vl1k2$nn9$525@news.insnet.com> Date: 19 Sep 1997 09:17:05 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.5vl1k2$nn9$525@news.insnet.com> Sender: goforit@savetrees.com Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: Eren_Kotan@apple.com (Eren Kotan) Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: Re: prelude to Rhapsody on Sparc?!?! Date: 19 Sep 1997 10:00:20 GMT Organization: NeXT/Apple European Professional Services Message-ID: <5vtifk$eei$1@news2.apple.com> References: <341E6923.2781E494@gatwick.geco-prakla.slb.com> <341E6990.446B9B3D@gatwick.geco-prakla.slb.com> <EGnt7v.2rL@novice.uwaterloo.ca> <3421263C.148F@emory.org> Cc: curtis_crowson@emory.org Various people wrote: > Have I correctely understood??? Prelude to Rhapsody is available on Sun > Sparc!!!??? "Prelude to Rhapsody" was simply a pre-release version of NeXT's OPENSTEP 4.2 product, intended to give developers a head start on getting the hang of the OpenStep APIs and development environment. OPENSTEP 4.2 supports intel-PCs, m68k-NeXTs and Sun-SPARC platforms. Don't confuse this with Rhapsody. > How comes? Will Rhapsody be available on Sparc? Apple has no announced plans to port Rhapsody to SPARC at this time. > Check out www.macosrumors.com, they mention that some internal documents > about Rhapsody mention bugs in drivers to other platforms like Sun, and > HP. A lot of the claims on www.macosrumors.com is incorrect. This is one of them. They must have heard of Portable Distributed Objecs (PDO) on Solaris and HP/UX and jumped to the conclusion that this is Rhapsody for SPARC and HP. It is NOT. PDO has been available for these two OSes for a long time, it is merely a collection of OpenStep kits, the NeXT compiler/debugger plus other assorted CLI devtools. It allows you to compile and deploy non-GUI OpenStep apps on these platforms (CLI server apps). There is no window server, no GUI devtools, and no AppKit. There is no support for GUI OpenStep apps. Hope this clears some of the confusion, Eren -- Eren Kotan The best friend money can buy. Can you imagine a world with no hypothetical questions? Apple Enterprise Professional Services - (formerly part of NeXT Software)
From: holger@object-factory.REMOVETHIS.com (Holger Hoffstaette) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 19 Sep 1997 09:33:34 GMT Organization: Object Factory GmbH (Germany) Message-ID: <5vtgte$h9c$1@leonie.object-factory.com> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> <5vqu4c$srg$1@leonie.object-factory.com> <3421B435.704@ibm.net> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Ralph Paul wrote: > Sorry this sounds like you have not seen a real flight control or > avionics system. While flight control system might be the result Personally? No. I do know people who are very involved with critical embedded systems in various ways - systems that kill people when they fail. One of them does things used in airplanes - admittedly, more targeted for the military markets. Also, I was not only referring to the stuff in the airplane - there's something on the ground, too. :) > Also you have to keep in mind that avionics or flight control system > have to be proven to work correctly not just by doing runtime tests but > on paper. Therefore Obj-C or C++ or Smalltalk or Eiffel are rather > unlikely options. Ada83 is commonly used due to it's rigid static > typing. Yes, and because a system has been *proven* to be correct it Just Can't Fail. I have it on paper! Whee.. (sorry, but formal verification doesn't fly with me) > Final comment : The flight control example from OOSC 2 > ( dynamic <-> static typing ) is really a bad one > because avionics, flight control systems are developed in a > very different manner than normal applications. > B. Meyer should have made his comments using a different example True. ObThread: BM was recently harshly criticized for his now-famous posting to comp.object: 'Avoiding the second historic mistake'. I thought it was one of BM's bright moments, and that he was right on target. Needless to say, he was blamed to be the Party Pooper Of The Day, spoiling all the Java hype. The funny thing is that some people over in comp.lang.smalltalk are already reporting the first defections from Java back to Smalltalk because nothing works, and Java's oh-so-small RAM footprint ain't so small anymore when you end up with hundreds useless little classes, all different. Interesting times, for sure.. Holger -- holger"at"object-factory.com (NeXTMail, MIME) 'Where I come from it's important to develop stuff that doesn't suck.' -- David Young <dwy"at"ace.net>
From: Chris Van Buskirk <cvbuskirk@home.com> Newsgroups: comp.sys.next.programmer Subject: its the 24th...... Date: Fri, 19 Sep 1997 08:33:55 -0400 Organization: NextGen Internet Message-ID: <34227133.37ED6D94@home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I read a macintouch, that the 24th is the day rhap will be released. cool... -chris
From: frank@this.NO_SPAM.net (Frank M. Siegert) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 19 Sep 1997 15:25:55 GMT Organization: Frank's Area 51 Message-ID: <5vu5i3$gqk$1@news.seicom.net> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> <5vqu4c$srg$1@leonie.object-factory.com> <3421B435.704@ibm.net> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: repaul@ibm.net In <3421B435.704@ibm.net> Ralph Paul wrote: > Sorry this sounds like you have not seen a real flight control or > avionics system. While flight control system might be the result > of many years of development, they certainly do not contain any > code that unnecessary since runtime performance is cirtical and > processor resources are always something that's too small. > Also you have to keep in mind that avionics or flight control system > have to be proven to work correctly not just by doing runtime tests but > on paper. Therefore Obj-C or C++ or Smalltalk or Eiffel are rather > unlikely > options. Ada83 is commonly used due to it's rigid static typing. > The rise and fall of the first launch of Ariane 5 proves otherwise. The code that caused the catastrophic events were left over from a previous release (for Ariane 4) but were in fact unneccessary for functionality. Surely there was a 'paper' test of this system before... -- * Frank M. Siegert [frank@this.net] - Home http://www.this.net * NeXTSTEP, Linux, BeOS & PostScript Guy
From: "Robert Fisher" <rfisher@onr.com> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 19 Sep 1997 15:34:45 GMT Organization: Renegade Software Message-ID: <01bcc474$1722d360$0100000a@ivanova> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> Robert A. Decker <comrade@umich.edu> wrote in article <B042DB4F-78B2@141.214.128.36>... > You forgot one of the most important things: Not having to worry about > type in Obj-C. After programming in Obj-C and Newtonscript and a little > Dylan I absolutely cannot stand the static typing of Java/C++. > In fact, static typing ruins the concept of OO programming. Why should I > care what type the object is that I just got? I should just be able to tell > it to do something and leave it at that. (that being said, I usually use > static typing in Obj-C except for a few functions). OK, let's see if I understand this. In Java, I have to cast an object to a specific class before I can call a method of that class with that object: (I'll ignore the fact that this code may throw an exception.) Object object; Frame frame; object = getSomeObject(); frame = (Frame) object; frame.show(); In Objective-C, I would've been able to do something like: (I apologize for any egregious mistakes. I haven't done any Objective-C programming yet.) id object; object = [self getSomeObject]; [object show]; I'm I understanding the point? -- Robert Fisher Renegade Software rfisher@onr.com
From: "Nate McNamara" <nate@team-sys.com> Newsgroups: comp.lang.c++,comp.sys.next.programmer Subject: Re: C++ RTTI vs Objective-C Date: 19 Sep 1997 16:26:07 GMT Organization: TEAM Systems Message-ID: <01bcc518$5a59e8e0$68646464@nate> References: <341EA2F3.3F54BC7E@gatwick.geco-prakla.slb.com> <rayEGMIEM.6nA@netcom.com> Ray, >>In Objective-C, the method called for a particular object can be changed at runtime.<< You can do this in C++ as well, although the details are rather nasty. See Coplien's _Advanced_C++_. Nate
From: rainer@mathematik.uni-wuerzburg.de (Rainer =?ISO-8859-1?Q?Frohnh=F6fer)?= Newsgroups: comp.sys.next.programmer,comp.sys.next.misc Subject: Re: prelude to Rhapsody on Sparc?!?! Date: 19 Sep 1997 18:52:38 GMT Organization: University of =?ISO-8859-1?Q?W=FCrzburg,?= Germany Message-ID: <5vuhlm$dm@lobotomy.urz.uni-wuerzburg.de> References: <341E6923.2781E494@gatwick.geco-prakla.slb.com> <341E6990.446B9B3D@gatwick.geco-prakla.slb.com> <EGnt7v.2rL@novice.uwaterloo.ca> <3421263C.148F@emory.org> <5vtifk$eei$1@news2.apple.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: Eren_Kotan@apple.com In <5vtifk$eei$1@news2.apple.com> Eren Kotan wrote: > > A lot of the claims on www.macosrumors.com is incorrect. This is one of them. > They must have heard of Portable Distributed Objecs (PDO) on Solaris and > HP/UX and jumped to the conclusion that this is Rhapsody for SPARC and HP. It > is NOT. PDO has been available for these two OSes for a long time, it is > merely a collection of OpenStep kits, the NeXT compiler/debugger plus other > assorted CLI devtools. It allows you to compile and deploy non-GUI OpenStep > apps on these platforms (CLI server apps). There is no window server, no GUI > devtools, and no AppKit. There is no support for GUI OpenStep apps. Plus, Sun doesn't seem too enthusiastic about OpenStep anymore. They killed OpenStep/Solaris. Probably considered it a threat to their Java thing. Stupid decision, now SPARCs aren't an alternative for me anymore. ------------------------------------- "Um Energie zu sparen, wird das Licht am Ende des Tunnels vorlaeufig abgeschaltet." rainer@mathematik.uni-wuerzburg.de (public key avaible at any key server near you ...)
From: ians@cam-ani.co.uk (Ian Stephenson) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: Fri, 19 Sep 1997 17:08:50 GMT Organization: Cambridge Animation Systems Ltd Sender: news@cam-ani.co.uk Message-ID: <EGrMAr.n30@cam-ani.co.uk> References: <01bcc474$1722d360$0100000a@ivanova> In article <01bcc474$1722d360$0100000a@ivanova> "Robert Fisher" <rfisher@onr.com> writes: > OK, let's see if I understand this. > > In Java, I have to cast an object to a specific class before I can call a > method of that class with that object: > (I'll ignore the fact that this code may throw an exception.) > > Object object; > Frame frame; > object = getSomeObject(); > frame = (Frame) object; > frame.show(); > > In Objective-C, I would've been able to do something like: > (I apologize for any egregious mistakes. I haven't done any Objective-C > programming yet.) > > id object; > object = [self getSomeObject]; > [object show]; > Pretty much. More signifigantly you can get into the state: Object object; object = getShowableObject(); if(object.iskindof(Frame)) { Frame frame; frame = (Frame) object; frame.show(); } else { Window window; window = (Window) object; window.show(); } Where you know the object responds to show, but don't have a common implementation/base class. The C++/Java version will break When I write a button class which also implements show. If works ok if I have a GuiObject class which implements a virtual show, but thats not always the case. $an $an
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <19970919184300.OAA14940@ladder01.news.aol.com> Control: cancel <19970919184300.OAA14940@ladder01.news.aol.com> Date: 19 Sep 1997 18:43:54 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.19970919184300.OAA14940@ladder01.news.aol.com> Sender: pptohealth@aol.com (Pptohealth) Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: mpaque.spa-am@nospam.wco.com (Mike Paquette) Newsgroups: comp.sys.next.advocacy,comp.sys.next.programmer Subject: Re: Rhapsody and SMP Followup-To: comp.sys.next.advocacy Date: Fri, 19 Sep 1997 19:05:40 GMT Organization: Electronics Service Unit No. 16 Message-ID: <3422ccb4.16657141@news.wco.com> References: <pxpst2-1809971432280001@pelli.pathology.pitt.edu> On Thu, 18 Sep 1997 14:32:28 -0400, pxpst2@vms.cis.pitt.edu (Peter) wrote: >I was hoping that some knowledgable peaple could explain how difficult the >task of implementing SMP in Rhapsody? It's pretty simple, really. 1) Get accurate documentation on the target SMP hardware. (Trickier than it looks in the Intel hardware space!) 2) Write a handfull of kernel code describing the hardware in terms of processor sets, thread migration, interlocks, and cache coherency management. 3) Set NCPUS to something greater than 1 and rebuild the kernel 4) Test. Then test some more. Then test again. (This bit takes some time...)
From: blenko@cs.yale.edu (Tom M. Blenko) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C?... categories contra OOP ?? Date: 19 Sep 1997 21:29:24 -0400 Organization: Computer Science Dept., Yale University Distribution: world Message-ID: <5vv8tk$neg@ruebezahl.cs.yale.edu> |> Also you have to keep in mind that avionics or flight control system |> have to be proven to work correctly not just by doing runtime tests but |> on paper. Therefore Obj-C or C++ or Smalltalk or Eiffel are rather |> unlikely options. Ada83 is commonly used due to it's rigid static |> typing. | |Yes, and because a system has been *proven* to be correct it Just |Can't Fail. I have it on paper! Whee.. |(sorry, but formal verification doesn't fly with me) Gee I hate to see an all-too-rare interesting and informative net discussion turn into an all-too-familiar flame fest. That's the sort of thing that drives sensible people away. It is true that "proving correctness" is a misnomer but Erik wasn't using it an overly strong way. He was talking about it as one technique, among several, used in practice to increase the reliability of programs. And he was quite careful about saying that its domain of applicability/desirability is limited. Tom
From: zizi zhao <ziziz@worldnet.att.net> Newsgroups: comp.sys.next.programmer Subject: Re: Replacing hard disk on Intel NS 3.2 Date: Sat, 20 Sep 1997 15:06:11 -0400 Organization: personal Message-ID: <6013iu$j1q@bgtnsc03.worldnet.att.net> References: <341AC34E.159E@decide.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: bobk@decide.com I just did duplicating on my system. I had an old disk, which is too small. I got a bigger one and did the following: 1. install the hard disk in my pc and boot nextstep; 2. initialize the new disk in nextstep format; 3. mount the new disk as /Disk; 4. (dump 0f - /dev/rold-disk) | (cd /Disk; restore -); 5. change your system Configuration to make sure you have driver for the new disk, if there wasn't one; 6. shutdown the system and take out the old disk; 7. boot; done. to be safe, i need double check my note in my office. zizi _____________________________________________________________ Bob Korsan wrote: > > Hi, > > I have had a NS 3.2 system for several years now (I'm not a developer, > just a Mathematica User and NS seemed to be the best OS). The disk drive > is now making wierd noises and I want to replace it before it crashes. I > have a 2 MB DAT drive which does a backup every night, so the info > should be there... If I go out and buy a new drive and install it, what > is the procedure to get 3.2 up and running again? The hardware vendor > did the initial install, so I have never done anything except run the > box. (Oh yes, this was an educational version of the OS so I have only > the most minimal of documentation -- I was teaching but have since > stopped. I don't have help from the campus computer support like I did > before.) > > Can anyone lay out the steps that I would have to follow? > Can anyone point me to a web site or ftp site which would have the info > I would like? (I tried NeXTanswers, but after about 2-3 hours of > wandering and searching that site, I couldn't find anything. I sent a > message to Apple Enterprise Support, but they say the OS is no longer > supported. They recommended I message this group, so I am.) > > Lastly, I haven't had any trouble with the OS, so I don't see the need > to upgrade, but if there is something really wiz bang on the latest > version, could someone let me know what I would be getting? And if you > know the price to upgrade from 3.2 to the latest (I don't even know what > the latest version is), please let me know. > > Thanks in advance for your help. > Peace/Bob
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <6015at$8b4603@volta.nmhu.edu> Control: cancel <6015at$8b4603@volta.nmhu.edu> Date: 20 Sep 1997 19:54:19 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.6015at$8b4603@volta.nmhu.edu> Sender: Andy<ss21@usa.net> Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: Bob Hathaway <bhathaway@sigs.com> Newsgroups: comp.sys.next.programmer,comp.lang.oberon,comp.object.logic,comp.lang.scheme,comp.lang.beta Subject: Object Magazine Online - CALL FOR PAPERS Followup-To: comp.object Date: 20 Sep 1997 21:41:42 GMT Organization: Object Magazine Online Distribution: world Message-ID: <601fum$tip$5@news.inch.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 8bit Summary: Object Magazine Online Call For Papers Keywords: Free WWW OO Object-Oriented Journal OBJECT MAGAZINE ONLINE - Call For Papers ========================================= MONTHLY OBJECT-ORIENTED FORUM FORMERLY: OBJECT CURRENTS http://www.sigs.com/omo (formerly Object Currents) Contact: bhathaway@sigs.com - Accepting articles on Object Technology - Articles can include Java, CORBA, DCOM, ActiveX, VRML, CGI, ... This is a call for contributors for the Online Edition of SIGS' Object Magazine. This Online Web Journal provides the opportunity to publish your work, with pay, on topics of interest to the community. Preferred features will make use of advanced object technology through the Web and include distributed objects, interaction, animation, and information technology. These can be implemented with CORBA, VRML, Java, Active-X, object databases, and etc. We should have an Iona Orb and the O2 object database available soon for use in articles and advanced arrangements for features will be discussed. All articles on object technology are also accepted. Object Currents has had some of the best world-class columnists and contributors over the past year and Object Magazine carries on the tradition. We'd also like to add more running examples and provide a forum for the distribution and discussion of new object technologies. We will have more to say on this subject soon and this is an invitation to join in bringing on the next wave of object technology to the forefront. Please take a look at the journal and see what it is all about: http://www.sigs.com/omo Thank You & Best Regards, Bob Hathaway Robert John Hathaway III Editor-In-Chief Object Magazine Online bhathaway@sigs.com http://www.sigs.com/omo
Newsgroups: comp.lang.c++,comp.sys.next.programmer From: ray@netcom.com (Ray Fischer) Subject: Re: C++ RTTI vs Objective-C Message-ID: <rayEGrx66.CKE@netcom.com> Organization: NETCOM, San Jose References: <341EA2F3.3F54BC7E@gatwick.geco-prakla.slb.com> <rayEGMIEM.6nA@netcom.com> <01bcc518$5a59e8e0$68646464@nate> Date: Fri, 19 Sep 1997 21:03:41 GMT Sender: ray@netcom6.netcom.com Nate McNamara <nate@team-sys.com> wrote: >>>In Objective-C, the method called for a particular object can be changed >at runtime.<< > > You can do this in C++ as well, although the details are rather nasty. >See Coplien's _Advanced_C++_. Alright. Yes, you really can do this in C++. YOu can do almost anything in C++. But the language does not really support altering the bindings at runtime like Objective-C does. And the resulting C++ code is probably going to be ugly and hard to maintain. -- Ray Fischer ray@netcom.com
From: Ingo Feulner <ifeulner@xenon.cube.de> Newsgroups: comp.sys.next.programmer Subject: Re: Api for Edit.app Date: 20 Sep 1997 08:56:59 GMT Organization: Private Site, =?ISO-8859-1?Q?B=F6blingen?=, Germany Message-ID: <60034r$2a8$1@xenon.cube.de> References: <5vo0cp$1ci@voyager.north.de> <EGo419.L1D@basil.icce.rug.nl> Mime-Version: 1.0 (FinalNews for OpenStep; Version 0.37 / Sep 14, 1997) Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Xcanpos: shelf.01/199709240201!0002776603 Tom Hageman wrote in <EGo419.L1D@basil.icce.rug.nl> on 1997-09-18 19:41:40 +0200: > The Speaker and Listener classes are only available in NeXTSTEP, not in > OPENSTEP. (Well, that's not entirely true. They are still declared in > <AppKit/obsoleteSpeaker.h> and <AppKit/obsoleteSpeaker.h>, at least in > OS/Mach.) Yes, there are still the headers, but it seems that Speaker/Listener does not work anymore... at least there are problems in communication with an app. I'm working on a project which needs to communicate with Mail.app; it worked under 3.3 but not anymore under OS4.2mach. Running the app in the debugger I can see that NXPortFromName succeeds but that the commands I send then to Mail.app do nothing. Any suggestions? -Ingo. -- Smail: Ingo Feulner, Wolfacher Weg 19, 71034 Böblingen, Germany Email: ifeulner@xenon.cube.de "If you want to convince yourself how nice OpenStep is, try Java" - Wiliam Shipley.
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <6026go$lud@new1.sundial.net> Control: cancel <6026go$lud@new1.sundial.net> Date: 21 Sep 1997 04:10:40 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.6026go$lud@new1.sundial.net> Sender: SEX APPEAL INSTANTLY<jason@compuserve.com> Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: sanguish@digifix.com Newsgroups: comp.sys.next.advocacy,comp.sys.next.announce,comp.sys.next.hardware,comp.sys.next.software,comp.sys.next.misc,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.sys.next.programmer Subject: NEXTSTEP/OpenStep Resources on the Net Supersedes: <9488874209621@digifix.com> Date: 21 Sep 1997 03:49:30 GMT Organization: Digital Fix Development Message-ID: <12335874814429@digifix.com> Topics include: Major OpenStep/NEXTSTEP World Wide Web Sites OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups Major OpenStep/NEXTSTEP FTP sites NeXTanswers Major OpenStep/NEXTSTEP World Wide Web Sites ============================================ The following sites are a sample of the OpenStep related WWW sites available. A comprehensive list is available on Stepwise. Stepwise OpenStep/Rhapsody Information Server http://www.stepwise.com Stepwise has been serving the OpenStep/NEXTSTEP community since March 1993. Some of the many resources on the site include: OpenStep Third Party Software guide, Developer Directory, Mailing List information, extensive listing of FTP and WWW sites related to OpenStep and NEXTSTEP, OpenStep related Frequently Asked Questions. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. This is the World Wide Web interace to the FTP site. Apple Enterprise Software Group (formerly NeXT Computer, Inc.) http://www.next.com Here is where you'll find the NeXTanswers archive, with information on OpenStep installation, drivers and software patches. Apple Computer's 'Prelude to Rhapsody' Self Support Site http://devworld.apple.com/dev/prelude.html This site has been constructed to help you help yourself to learn as much as possible about the foundation for Rhapsody, today's OPENSTEP. The site provides an informal collection of pointers, references, and starting points for developers who are using the Prelude to Rhapsody bundle, distributed at this year's Worldwide Developer Conference. OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups ==================================================== COMP.SYS.NEXT.ADVOCACY This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. COMP.SYS.NEXT.ANNOUNCE Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. COMP.SYS.NEXT.BUGS A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT-specific groups as well. COMP.SYS.NEXT.HARDWARE Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. COMP.SYS.NEXT.MARKETPLACE NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. COMP.SYS.NEXT.MISC For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! COMP.SYS.NEXT.PROGRAMMER Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. COMP.SYS.NEXT.SOFTWARE This is a place to talk about [third party] software products that run on NEXTSTEP systems. COMP.SYS.NEXT.SYSADMIN Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. ** RELATED NEWSGROUPS ** COMP.SOFT-SYS.NEXTSTEP Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. COMP.LANG.OBJECTIVE-C Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. COMP.OBJECT Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep next-advocacy next-announce next-bugs next-hardware next-marketplace next-misc next-programmer next-software next-sysadmin object lang-objective-c (For a full description, send mail to listserv@antigone.com). The subscription syntax is essentially the same as Majordomo's. To subscribe, send a message to *-request@lists.best.com saying: subscribe where * is the name of the list e.g. next-programmer-request@lists.best.com Major OpenStep/NEXTSTEP FTP sites ================================= ftp://ftp.next.peak.org The main site for North American submissions formerly ftp.cs.orst.edu ftp://ftp.peanuts.org: (Peanuts) Located in Germany. Comprehensive archive site. Very well maintained. ftp://ftp.nluug.nl/pub/comp/next NeGeN/NiNe (NEXTSTEP Gebruikers Nederland/NeXTSTEP in the Netherlands) ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! _________________________________________________________________ Written by: Eric P. Scott ( eps@toaster.SFSU.EDU ) and Scott Anguish ( sanguish@digifix.com ) Additions from: Greg Anderson ( Greg_Anderson@afs.com ) Michael Pizolato ( alf@epix.net ) Dan Grillo ( dan_grillo@next.com )
From: patrickl@aei.ca (Patrick Loutfi) Newsgroups: comp.sys.next.programmer Subject: Wanted Debugging or Modification Contract Date: Sun, 21 Sep 1997 04:04:07 GMT Organization: AEI Message-ID: <34249ca5.3001656@news.aei.ca> Wanted debugging or modification contract Do you need a hand debugging or modifying a program written in Basic, Visual Basic or C ? Well maybe I can help you with that. I’m a Microsoft certified Visual Basic (Exam 70-65) programmer with 2 years experience in basic. I started my own company 2 years ago in an other field, but my true love is programming which is why I am offering my services. I have great debugging skills and my rates are very reasonable. Reply at patrickl@aei.ca Thanks
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <34249ca5.3001656@news.aei.ca> Control: cancel <34249ca5.3001656@news.aei.ca> Date: 21 Sep 1997 05:15:55 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.34249ca5.3001656@news.aei.ca> Sender: patrickl@aei.ca (Patrick Loutfi) Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <34250334.8@news1.abac.com> Control: cancel <34250334.8@news1.abac.com> Date: 21 Sep 1997 11:17:09 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.34250334.8@news1.abac.com> Sender: Me Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.lang.objective-c,comp.sys.next.programmer Subject: RFC: The comp.lang.Objective-C FAQ Followup-To: comp.lang.objective-c Date: 21 Sep 1997 19:02:40 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7vhzufvm7.fsf@tom.ics.ele.tue.nl> Part 1 of this month's FAQ (the `Answers') has seen major edits, especially in questions 1-9 and 14-16. Comments are welcome. --Tiggr
From: steve@spvi.com (Steve Spicklemire) Newsgroups: comp.sys.next.programmer Subject: Rhapsody RDR1 Date: 21 Sep 1997 19:51:32 GMT Organization: IndyNet - Indys Internet Gateway (info@indy.net) Message-ID: <603ts4$268$1@news.indy.net> Hi there, Does anyone know or suspect that code compiled under Prelude will ahve any hope running unchanged (e.g., not recompiled) under RDR/Intel? -steve
From: kaared@hiMolde.no (Kare Digernes) Newsgroups: comp.sys.next.programmer Subject: Any books on NeXTSTEP and OpenStep programming? Date: 21 Sep 1997 20:19:31 GMT Organization: UNINETT news service Message-ID: <874873170.986004@dole.uninett.no> Cache-Post-Path: dole.uninett.no!unknown@ulke.himolde.no I am looking for recommendations on books that teach programming for NeXTSTEP and OpenStep. Thank you in advance for any input. -- -Kaare kaared@himolde.no
Message-ID: <34258BD9.6B32@ibm.net> Date: Sun, 21 Sep 1997 22:04:26 +0100 From: Ralph Paul <repaul@ibm.net> MIME-Version: 1.0 Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> <5vqu4c$srg$1@leonie.object-factory.com> <3421B435.704@ibm.net> <5vu5i3$gqk$1@news.seicom.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Frank M. Siegert wrote: > The rise and fall of the first launch of Ariane 5 proves otherwise. The code > that caused the catastrophic events were left over from a previous release > (for Ariane 4) but were in fact unneccessary for functionality. Surely there > was a 'paper' test of this system before... True, but for whatever reason they did perform a functional benchmark test with the Ariane 5 trajectory. For more on this see the language debatte that raged for many weeks on comp.lang.ada, com.lang.eiffel on this. I hope we can stop this discussion right here ? Ralph Paul
From: Josh Sahrmann <Jsahrman@cstone.net> Newsgroups: comp.sys.next.programmer Subject: Speech Date: Sun, 21 Sep 1997 17:33:13 +0000 Organization: Gates & Reasons Message-ID: <34255A49.91D5B59E@cstone.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Rhapsody is being sent to me and im trying to work out a software plan but for software that I have to work out but im not sure if Rhapsody has Speech. For the people who are reading this could you tell me if Rhapsody has Speech like MacOS has. The idea is to add Speech effects to the software. -- Josh Sahrmann Macintosh Developer Gates & Reasons Inc. Be Developer
From: Steve Dekorte <dekorte@slip.net> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 22 Sep 1997 02:24:57 GMT Organization: Slip.Net (http://www.slip.net) Message-ID: <604ktp$m47$2@owl.slip.net> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <01bcc474$1722d360$0100000a@ivanova> Robert Fisher <rfisher@onr.com> wrote: > Robert A. Decker <comrade@umich.edu> wrote in article > <B042DB4F-78B2@141.214.128.36>... > > In fact, static typing ruins the concept of OO programming. Why should > OK, let's see if I understand this. [example of simple hard-coded message send deleted] Ok, now try to implement a Collection class that can hold mixed objects. And a class that can send any message to any object at runtime. And an example of sending a message to an object that doesn't respond to it but can forward it would be neat. -- Steve Dekorte - OpenStep consultant - San Francisco
From: sjf@ids.net Newsgroups: comp.sys.next.programmer Subject: Q: NeXTStep -> SCO UNIX Conversion Date: Mon, 22 Sep 1997 02:17:10 GMT Organization: IDS World Network Internet Access Service, (401) 885-4243 Message-ID: <3425d367.2024794@paperboy.ids.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello I don't much experience using NeXTStep but, I was wondering if anyone has any experience porting programs which run on NeXTStep to SCO UNIX. I'm assuming that since, NeXTStep is very close to a flavor of UNIX as long as the program uses POSIX Standard scripting I should be ok. Any Comments? Steven J. Franko Mail: sjf@ids.net. sjf@mcprx.com, SJFranko@cvs.com URL: http://users.ids.net/~sjf
From: Steve Dekorte <dekorte@slip.net> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 22 Sep 1997 02:14:08 GMT Organization: Slip.Net (http://www.slip.net) Message-ID: <604k9g$m47$1@owl.slip.net> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> <5vrall$ghn1@castor.cca.rockwell.com> > For example, the Objective-C message dispatch could probably never be proven > for use in airplanes because it uses (difficult to prove) hash tables and > dynamic allocation. Uh, this may be a stupid question, but is it possible to write and Objective-C interpreter in ADA? If so how provable would what it interpret's be? Extend these thoughts to "provability" problem in general. -- Steve Dekorte - OpenStep consultant - San Francisco
From: goforit@savetrees.com Newsgroups: comp.sys.next.programmer Subject: What if . . . Date: 16 Sep 1997 04:23:30 GMT Organization: The Alliance Message-ID: <5vl1k2$nn9$525@news.insnet.com> Ten years ago, you could have bought 36 shares of Microsoft for 28.01 per share. That $1,008.36 would be $198,126.00 now. What if you had another chance to invest in a company with that much potential? If you leave your email at "http://theprofitcenter.net/stock" you could get the chance to consider an Internet investment with tremendous potential.
From: seanlNoSpam@carmi.cs.umd.edu (Sean Luke) Newsgroups: comp.lang.c++,comp.sys.next.programmer Subject: Re: C++ RTTI vs Objective-C Followup-To: comp.lang.c++,comp.sys.next.programmer Date: 22 Sep 1997 15:31:40 GMT Organization: University of Maryland at College Park Message-ID: <60630s$giv$1@walter.cs.umd.edu> References: <341EA2F3.3F54BC7E@gatwick.geco-prakla.slb.com> <rayEGMIEM.6nA@netcom.com> <01bcc518$5a59e8e0$68646464@nate> <rayEGrx66.CKE@netcom.com> Ray Fischer (ray@netcom.com) wrote: >Nate McNamara <nate@team-sys.com> wrote: >>>>In Objective-C, the method called for a particular object can be changed >>at runtime.<< >> >> You can do this in C++ as well, although the details are rather nasty. >>See Coplien's _Advanced_C++_. > >Alright. Yes, you really can do this in C++. YOu can do almost >anything in C++. But... Yep. If a language wasn't turing-complete, no one would use it (okay, except maybe HyperTalk, right Lawson? :-) The point of language design isn't what *can* be done with a language, but how well a language supports specific mechanisms in terms of codability, readability, resource utilization, and maintainability. Otherwise, we'd all just code in assembly. What a language "can" (so to speak) do is a really dumb argument. _____________________________________________________________________________ Sean Luke Spam Must Die! "I've discovered that P==NP, but the proof is too U Maryland at College Park large to fit in the margins of this signature." seanl@nospamcs.umd.edu URL: http://nospamwww.cs.umd.edu/~seanl/
From: embuck@palmer.cca.rockwell.com (Erik M. Buck) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 22 Sep 1997 15:52:19 GMT Organization: Rockwell Collins Message-ID: <60647j$mj1@castor.cca.rockwell.com> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> <5vrall$ghn1@castor.cca.rockwell.com> <604k9g$m47$1@owl.slip.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: dekorte@slip.net In <604k9g$m47$1@owl.slip.net> Steve Dekorte wrote: > > For example, the Objective-C message dispatch could probably never be proven > > for use in airplanes because it uses (difficult to prove) hash tables and > > dynamic allocation. > > Uh, this may be a stupid question, but is it possible to write and Objective-C > interpreter in ADA? If so how provable would what it interpret's be? Extend these > thoughts to "provability" problem in general. > > It would not matter if it was implemented in machine language. It is the algorithm that is difficult to prove. The language is irrelevant
Message-ID: <34259877.2106@ibm.net> Date: Sun, 21 Sep 1997 22:58:16 +0100 From: Ralph Paul <repaul@ibm.net> MIME-Version: 1.0 Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> <5vqu4c$srg$1@leonie.object-factory.com> <3421B435.704@ibm.net> <5vtgte$h9c$1@leonie.object-factory.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Holger Hoffstaette wrote: [snip] > Also, I was not only referring to the stuff in the airplane - > there's something on the ground, too. :) True ! [snip] > Yes, and because a system has been *proven* to be correct it Just > Can't Fail. I have it on paper! Whee.. > (sorry, but formal verification doesn't fly with me) > I didn't say that. I would be a big fool if I believed that. Formal proof and proving that a system is correct are not the same. Formal proof is part of the proof process but implementation details can still lead to unsafe software and have to be checked. This is the reason why flight control software is tested using several different steps : formal proof, non realtime simulation, realtime simulation, hardware in the loop sims, hardware in the loop + rig testing I am sure your friends can tell you more about the subject of software classification, quality assurance, testing, ...... [snip] > ObThread: > BM was recently harshly criticized for his now-famous posting to > comp.object: 'Avoiding the second historic mistake'. I thought it > was one of BM's bright moments, and that he was right on target. > Needless to say, he was blamed to be the Party Pooper Of The Day, > spoiling all the Java hype. Took at look at the original post on dejanews. We don't have to repeat it (:-). > The funny thing is that some people over in comp.lang.smalltalk are > already reporting the first defections from Java back to Smalltalk > because nothing works, and Java's oh-so-small RAM footprint ain't > so small anymore when you end up with hundreds useless little > classes, all different. > > Interesting times, for sure.. Indeed. It seems the Java language has peaked on the hype scale and is now on the way to become just another languages. Even Apple has delayed the Java bindings for rhapsody. I found the iX Magazin editorial quite interesting. Servus, Ralph Paul
From: Curtis Crowson <curtis_crowson@emory.org> Newsgroups: comp.lang.c++,comp.sys.next.programmer Subject: Re: C++ RTTI vs Objective-C Date: Mon, 22 Sep 1997 13:29:14 -0400 Organization: Emory University System of Health Care Message-ID: <3426AAEA.1E1D@emory.org> References: <341EA2F3.3F54BC7E@gatwick.geco-prakla.slb.com> <rayEGMIEM.6nA@netcom.com> <01bcc518$5a59e8e0$68646464@nate> <rayEGrx66.CKE@netcom.com> <60630s$giv$1@walter.cs.umd.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sean Luke wrote: > Yep. If a language wasn't turing-complete, no one would use it (okay, > except maybe HyperTalk, right Lawson? :-) Because turing-complete involves an infinite storage device, I suspect that no language is turing complete. Except for the infinite storage part of a turing machine all you need is a subtract and jump on minus instruction to supposedly be turing complete. I would hate to code in this language though. And don't tease Lawson since he has been banished. :-) Its just not nice.
From: jason@fisher.psych.uh.edu (Jason L. Asbahr) Newsgroups: comp.lang.python,comp.sys.next.programmer Subject: NeXT-style Serializer for Python? Date: 22 Sep 97 13:52:44 Organization: C.R.A.S.H. http://www.crash.org Message-ID: <JASON.97Sep22135244@fisher.psych.uh.edu> Hi! I want to exchange object data between Python apps and OPENSTEP apps. Python's pickle is close, but as it says in the docs, it's Python specific. In general, pickle seems analogous to NeXT's NSArchiver. I'm wondering if anyone has yet written a Python module that can serialize Dictionaries, Arrays, and Strings in a format compatible with the NeXT OPENSTEP NSSerializer? One use for this would be to read EOModel files into your Python programs, though why exactly you would want to do with is left as an exercise for the reader. ;-) Thanks, Jason Asbahr
From: John Kuszewski <johnk@spork.niddk.nih.gov> Newsgroups: comp.sys.next.programmer Subject: new dev camp? Date: Mon, 22 Sep 1997 16:22:45 -0400 Organization: Nat'l Insts of Health Message-ID: <3426D395.3882@spork.niddk.nih.gov> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, This appeared on Macintouch today: > Apple Europe is setting up a FREE training program (1 week) for > developers who wish to learn Rhapsody. Does anyone know anything more about this? I just sent in my official Apple Developer stuff a few days ago. Is this real? Are American developers eligible for an equivalent course? When's it starting? Is it like the old NeXT developer camps? -- _____________ | ___/_ | |/ / -- /\ // /-- || || / /|| || || / / || || ||/ / || John Kuszewski || |/ /| || johnk@spasm.niddk.nih.gov || / /|| || \/ / / || \/ that's MISTER protein G to you! |/__/| | /_________| My parents went to Zaire and all I got was this lousy retrovirus.
From: agnus@amylnd.s.bawue.de (Matthias Zepf) Newsgroups: comp.sys.next.programmer Subject: Initializing all NSLocks before becoming multithreaded Date: 22 Sep 1997 08:34:54 GMT Organization: NewsInDosen-Testbetrieb auf amylnd Message-ID: <605aje$fib$1@amylnd.s.bawue.de> Mime-Version: 1.0 (FinalNews for OpenStep; Version 0.37 / Sep 14, 1997) Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Xcanpos: shelf.02/199709260201!0000322225 Together with a friend I'm currently writing a highly multithreaded application for OpenStep. To synchronize parts of code and objects we use a lot of NSLocks (and NSConditionLocks). When reading the documentation on NSLock we found out that NeXT/Apple requests to initialize _ALL_ NSLocks before the application becomes multithreaded. Do anyone know why? It is important for us because in our application it is impossible to do so. When already become multithreaded we allocate a lot of objects that each requires one NSLock and we do not have any chance to known how much objects we will need before becoming multithreaded. In the current implementation we just alloc and init NSLocks when we need them and the application works that way. Matthias -- ** Matthias Zepf, Leonberg, Deutschland/Germany, http://www.bawue.de/~agnus ** ** E-Mail (NeXTmail/MIME): agnus@amylnd.s.bawue.de ** ** PGP key on request **
From: Steve Dekorte <dekorte@slip.net> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 22 Sep 1997 22:20:40 GMT Organization: Slip.Net (http://www.slip.net) Message-ID: <606qvo$855$1@owl.slip.net> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> <5vrall$ghn1@castor.cca.rockwell.com> <604k9g$m47$1@owl.slip.net> <60647j$mj1@castor.cca.rockwell.com> Erik M. Buck <embuck@palmer.cca.rockwell.com> wrote: > In <604k9g$m47$1@owl.slip.net> Steve Dekorte wrote: > > Uh, this may be a stupid question, but is it possible to write and > > Objective-C > > interpreter in ADA? If so how provable would what it interpret's be? Extend > > these thoughts to "provability" problem in general. > > thoughts to "provability" problem in general. > > > It would not matter if it was implemented in machine language. It is the > algorithm that is difficult to prove. The language is irrelevant Exactly. And if you need to do something dynamic, you either use well tested libraries (like the Objective-C runtime) or you write your own. -- Steve Dekorte - OpenStep consultant - San Francisco
From: Ingo Feulner <ifeulner@xenon.cube.de> Newsgroups: comp.sys.next.programmer Subject: Implementation behind NSLocks?? Date: 22 Sep 1997 21:05:16 GMT Organization: Private Site, =?ISO-8859-1?Q?B=F6blingen?=, Germany Message-ID: <606mic$pqf$1@xenon.cube.de> Mime-Version: 1.0 (FinalNews for OpenStep; Version 0.37 / Sep 14, 1997) Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Xcanpos: shelf.02/199709260201!0093563252 Hello, a friend and I are currently working on a heavily multi-threaded app (OS4.2) We're using NSLock, NSConditionLock and NSRecursiveLock instances to protect the data shared between the different threads. The problem is, that we have to alloc instances of these classes after the app has become multithreaded. But especially this seems not to be supported, as mentioned in the documentation of the NS*Lock classes. I ask now: why? This limitation does not exist for Mach mutex locks, and it does not exist for the objects wrappers for these locks implemented by the NX*Lock classes. Can anyone who knows more about the implementation of the NSLocks tell me why this limitation is there? What kind of workarounds do you suggest? (We cannot alloc a static pool of NSLock instances before the app becomes multithreaded because we do not know how many locks we need... it depends on the user actions...) Thanks, Ingo. -- Smail: Ingo Feulner, Wolfacher Weg 19, 71034 Böblingen, Germany Email: ifeulner@xenon.cube.de "If you want to convince yourself how nice OpenStep is, try Java" - Wiliam Shipley.
From: guyt@is.twi.tudelft.nl (A. Guyt) Newsgroups: comp.sys.next.programmer Subject: Re: Implementation behind NSLocks?? Date: 23 Sep 1997 08:04:58 GMT Organization: Delft University of Technology Message-ID: <607t7a$opu$1@news.tudelft.nl> References: <606mic$pqf$1@xenon.cube.de> Ingo Feulner <ifeulner@xenon.cube.de> writes > Hello, > > a friend and I are currently working on a heavily multi-threaded app (OS4.2) > We're using NSLock, NSConditionLock and NSRecursiveLock instances to protect > the data shared between the different threads. > > The problem is, that we have to alloc instances of these classes after the > app has become multithreaded. But especially this seems not to be supported, > as mentioned in the documentation of the NS*Lock classes. > I ask now: why? > This limitation does not exist for Mach mutex locks, and it does not exist > for the objects wrappers for these locks implemented by the NX*Lock classes. > This is really, really strange... It is simplly nuts to forbid the creation of locks after the creation of any thread !! One cannot write serious threaded apps with this restriction. Any object created at run time on the fly that needs to be thread safe needs its own lock !! Nor Mach nor Windows NT has this restriction of its own. It is even stranger that Apple's own documentation contradicts this restriction. Look at the NSlocking protocol documentation. The example included shows an object that calls a method creating a lock after the objects detect the multithreaded state ! Any Apple developer here to clarify this absurd situation ? Abraham Guyt. _____________________________________________________________________ Abraham Guyt P.O.Box 356 Department of Information Systems 2600 AJ Delft Faculty Information Technology & Systems The Netherlands Delft University of Technology tel: +31 15 278 5969 E-mail: guyt@is.twi.tudelft.nl NeXT-mail welcome
From: seanlNoSpam@carmi.cs.umd.edu (Sean Luke) Newsgroups: comp.lang.c++,comp.sys.next.programmer Subject: Re: C++ RTTI vs Objective-C Followup-To: comp.lang.c++,comp.sys.next.programmer Date: 23 Sep 1997 18:51:41 GMT Organization: University of Maryland at College Park Message-ID: <60933t$i98$1@walter.cs.umd.edu> References: <341EA2F3.3F54BC7E@gatwick.geco-prakla.slb.com> <rayEGMIEM.6nA@netcom.com> <01bcc518$5a59e8e0$68646464@nate> <rayEGrx66.CKE@netcom.com> <60630s$giv$1@walter.cs.umd.edu> <3426AAEA.1E1D@emory.org> Curtis Crowson (curtis_crowson@emory.org) wrote: >Because turing-complete involves an infinite storage device, I suspect >that no language is turing complete. This is somewhat of a misnomer, but *languages* are certainly turing-complete. Whether or not the machine they run on has infinite resources is another issue entirely. Gettin' off-topic, tho. _____________________________________________________________________________ Sean Luke Spam Must Die! "I've discovered that P==NP, but the proof is too U Maryland at College Park large to fit in the margins of this signature." seanl@nospamcs.umd.edu URL: http://nospamwww.cs.umd.edu/~seanl/
From: guyt@is.twi.tudelft.nl (A. Guyt) Newsgroups: comp.sys.next.programmer Subject: Answer to NSLock questions here Date: 24 Sep 1997 10:32:43 GMT Organization: Delft University of Technology Message-ID: <60aq8b$1q3$1@news.tudelft.nl> L.S., This morning I received an e-mail from one of Apple's senior Technical Writers in which he clarifies the NSLock policy as following: "A multithreaded application can create threadsafe objects at runtime, but it must create the lock when it creates the object and make this lock available to other objects in the application. More specifically, a multithreaded app should have a section of code that is only run by one thread (usually the main thread) and this section of code is _the_ place in the application where locks are created and handed off to objects elsewhere in the application." It seems that the requirement to create all locks before the app becoming multithreaded is a documentation error. My e-mail to Apple concerning this issue is being forwarded to the responsible class doc writer. Regards, Abraham Guyt. _____________________________________________________________________ Abraham Guyt P.O.Box 356 Department of Information Systems 2600 AJ Delft Faculty Information Technology & Systems The Netherlands Delft University of Technology tel: +31 15 278 5969 E-mail: guyt@is.twi.tudelft.nl NeXT-mail welcome
From: frank@this.NO_SPAM.net (Frank M. Siegert) Newsgroups: comp.sys.next.software,comp.sys.next.programmer,comp.lang.asm Subject: Re: assembler for PPro & NEXTSTEP? Date: 24 Sep 1997 11:57:25 GMT Organization: Frank's Area 51 Message-ID: <60av75$fma$1@news.seicom.net> References: <60aljn$n1f$1@lyra.csx.cam.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Cc: tph1001@CL.cam.ac.uk In <60aljn$n1f$1@lyra.csx.cam.ac.uk> Thomas Harte wrote: > > Is there an assembler (commercial or otherwise) available for > Intel PPro running NEXTSTEP 4.1? 'man as' AS(1) UNIX Programmer's Manual AS(1) NAME as - NeXT GNU-based assemblers producing Mach object files SYNOPSIS as [ option ... ] [ file ... ] DESCRIPTION The as command translates assembly code in the named files to object code. If no files are specified, as reads from stdin. All undefined symbols in the assembly are treated as global. The output of the assembly is left in the file a.out by default. ... That's on 3.3, guess it is still in 4.x... -- * Frank M. Siegert [frank@this.net] - Home http://www.this.net * NeXTSTEP, Linux, BeOS & PostScript Guy
From: mmalcolm crawford <malcolm@plsys.co.uk> Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 24 Sep 1997 13:18:25 GMT Organization: P&L Systems Message-ID: <60b3v1$4oe$1@ironhorse.plsys.co.uk> References: <5vds3m$di4$1@ironhorse.plsys.co.uk> <6096im$dfi$1@netty.york.ac.uk> In-Reply-To: <6096im$dfi$1@netty.york.ac.uk> On 09/23/97, -bat. wrote: >mmalcolm crawford <malcolm@plsys.co.uk> writes: >> If you are running OPENSTEP/4.x Mach go to the Termnial and do: >> >> defaults write NSGlobalDomain NSInterfaceStyle Windows95 > >...and then spend 3/4 of an hour trying to fine anything which has changed >in the slightest ! The only noticeable thing I could see was that the >scrollbars on Interface Builders window has changed. Apart from that >everything was identical. Am I missing something ? I expected to see >OpenStep applications looking more "windows 95"ish somehow... > I think maybe you need to restart relevant (4.2) applications... Try logging out then in again. If this fails, maybe a visit to SpecSaver...? :-) Best wishes, mmalc.
From: rog@ohm.york.ac.uk (Roger Peppe) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 24 Sep 1997 13:17:45 GMT Organization: Department of Electronics, University of York, UK. Sender: rp9@york.ac.uk Message-ID: <60b3tp$stk$1@netty.york.ac.uk> References: <5v2gar$8e@saturn.genoa.com> <5v3uet$qkf@shelob.afs.com> On 9 Sep 1997 16:45:17 GMT, Gregory H. Anderson <Greg_Anderson@afs.com> wrote: > I happen to agree with Alex's point of view. The one gotcha is that > _this_ construction: > > float value = [aField floatValue]; > > returns different values on different architectures when aField == nil. > On m68k, value = 0 after this message. In i386, it can be NaN. This is > clearly an undesirable result. So you do have to be careful about sending > messages to nil objects that return certain types of values. not only this, but something like: NSRect fr = [aView frame]; might easily crash the application with stack corruption if aView is nil. a large number of my more intractable bugs have been associated with bits of code failing silently because of an uninitialised variable somewhere. (particularly when instance variable connections have been lost in Interface Builder) i'd say that: if (foo) [foo bar]; is actually much better with the test than without. when reading the above code, it's obvious that the [foo bar] method might never be called. i see very little difference between nil objects and NULL pointers - IMO dereferencing either should cause a core dump. allowing calls to nil objects is just an excuse for programmer laziness... rog.
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 24 Sep 1997 15:21:14 GMT Organization: Anderson Financial Systems Inc. Message-ID: <60bb5a$a0p@shelob.afs.com> References: <60b3tp$stk$1@netty.york.ac.uk> Roger Peppe writes > Gregory H. Anderson <Greg_Anderson@afs.com> wrote: > > _this_ construction: > > > > float value = [aField floatValue]; > > > > returns different values on different architectures when aField == nil > > not only this, but something like: > > NSRect fr = [aView frame]; > > might crash the application with stack corruption if aView is nil. BTW, the reason for this problem is that when a method returns a struct, the compiler uses an alternate form of objc_msgSend() that passes the return value as a 3rd hidden argument on the stack (like self and _cmd). I'm sure it doesn't got to the trouble of initializing that memory, which can lead to unpredictable results if the method never gets called due to (aView == nil). Been there, done that. -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
From: dicosmo@verveine.ens.fr (Roberto Di Cosmo) Newsgroups: comp.sys.next.programmer Subject: Internet Phone for NeXT: partial port (if you know SoundLib you can help) Date: 24 Sep 1997 17:45:12 +0200 Organization: ENS Message-ID: <scu3fak96f.fsf@verveine.ens.fr> This is to announce a first trial to compile Speak-Freely 6.1 on a NeXTStep system. If you know how to program sound play and recording you can definitely help :-) I have chosen Sepak Freely (by John Walker) as it seems one of the better written programs, and because the source code is available, then modified it slightly to try and run under Nextstep... The code is available as : http://www.dmi.ens.fr/~dicosmo/SW/NeXT/sfnext.tar.gz The sound is still not really good, but I need help from somebody knowing bettern than me how play and record can be handled in NextStep (I strongly suspect that SoundLib cannot handle sequences short sound play/records properly. Follows the README file: // 18 September 1997 This first trial to compile speak-freely on a NeXT is not fully operational: - sound recording and sound playing functions behave very strangely on the .5 second sound sequences generated by speak-freely: you clearly hear a clicking noise and time separation - this behaviour is *not* my fault: I added some code that saves the bits of sound in consecutive files in /tmp/xxx.snd and the full final sounds in /tmp/log.snd You can check that the demo program chaintest (tried this on 3.3 HPPA and 3.3 Intel) has the same clicking noise over the /tmp/xxx.snd while the /tmp/log.snd is correct. (try sfspeaker & sfmike localhost ring.au) - if you know how to overcome this difficulty, let me know :-) - by the way, NextTime installs a special sound driver for probably the same reasons... Is there any documentation on its API that could help? - as far as I could verify on HPPA and Intel using SB16Pnp, no full duplex is supported, so HALF_DUPLEX is defined -- Roberto Di Cosmo -------------------------------------------------------------- LIENS-DMI E-mail: dicosmo@dmi.ens.fr Ecole Normale Superieure WWW : http://www.ens.fr/~dicosmo 45, Rue d'Ulm Tel : ++33-(0)1-44 32 20 40 75230 Paris CEDEX 05 Fax : ++33-(0)1-44 32 20 80 FRANCE MIME/NextMail accepted --------------------------------------------------------------
From: embuck@palmer.cca.rockwell.com (Erik M. Buck) Newsgroups: comp.sys.next.programmer Subject: Where to put frameworks Date: 24 Sep 1997 16:46:59 GMT Organization: Rockwell Collins Message-ID: <60bg63$6tj2@castor.cca.rockwell.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I have an application that uses frameworks. I would like to put the application on the intranet for others to use. Where should I put the frameworks ? As it is now, the frameworks are installed in $(HOME)/Library/Frameworks and and the application is installed in $(HOME)/Apps. If I change the framework installation to the network shared /LocalLibrary/Frameworks and rebuild the app to look for frameworks in /LocalLibrary/Frameworks, then I can no longer build new versions for my own use without corrupting the network version. The DYNL_LOADING_PATH environment variable is handy, but it only works within a shell (not when double click documents or the app) Where should I put the DLLs for the Windows NT users ?
From: "CALL-NOW !!!" email97@email97now3.com Newsgroups: comp.sys.next.programmer Subject: ¤ ¤ ¤ F R E E V A C A T I O N S !!! ¤ ¤ ¤ Date: Wed, 24 Sep 1997 15:58:34 -0700 Organization: GHIJ Message-ID: <240997155834@email97now3.com> ¤ ¤ ¤ F R E E V A C A T I O N S !!! ¤ ¤ ¤ 3 Day / 2 Night Vacation for up to a family of 5 http://www.vacationpromotions.com ><=<=>>=<<<<<>>>=<><
From: Steve Dekorte <dekorte@slip.net> Newsgroups: comp.sys.next.programmer Subject: Re: Where to put frameworks Date: 24 Sep 1997 23:39:35 GMT Organization: Slip.Net (http://www.slip.net) Message-ID: <60c8bn$l8p$1@owl.slip.net> References: <60bg63$6tj2@castor.cca.rockwell.com> Erik M. Buck <embuck@palmer.cca.rockwell.com> wrote: > I have an application that uses frameworks. > I would like to put the application on the intranet for others to use. > Where should I put the frameworks ? [various problems mentioned] All I want for Christmass is to be able to put frameworks in app wrappers. -- Steve Dekorte - OpenStep consultant - San Francisco
From: sef@kithrup.com Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <240997155834@email97now3.com> Date: 24 Sep 1997 22:08:44 GMT Control: cancel <240997155834@email97now3.com> Message-ID: <cancel.240997155834@email97now3.com> Sender: "CALL-NOW !!!" email97@email97now3.com Spam cancelled by sef@kithrup.com
From: "Jeffrey D. Iverson" <jeffiv@cwcinc.com> Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: Directory update Date: Wed, 24 Sep 1997 12:39:22 -0500 Organization: Iverson Software Co. Message-ID: <3429504A.3EDF@cwcinc.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit The Directory of Software Development Professionals (http://www.iversonsoftware.com/service.html) now has a category for NeXTStep/OpenStep. -- The Directory of Software Development Professionals http://www.iversonsoftware.com/service.html http://www.iversonsoftware.com/add_entry.htm
From: Thomas Harte <tph1001@CL.cam.ac.uk> Newsgroups: comp.sys.next.software,comp.sys.next.programmer,comp.lang.asm Subject: Re: assembler for PPro & NEXTSTEP? Date: 25 Sep 1997 09:08:54 GMT Organization: Cambridge University Message-ID: <60d9n6$703$1@lyra.csx.cam.ac.uk> References: <60aljn$n1f$1@lyra.csx.cam.ac.uk> <60av75$fma$1@news.seicom.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: frank@this.NO_SPAM.net In <60av75$fma$1@news.seicom.net> Frank M. Siegert wrote: > 'man as' > > AS(1) UNIX Programmer's Manual AS(1) > > NAME > as - NeXT GNU-based assemblers producing Mach object files > Sorry. I should have added that I last programmed an 8086 using Turbo Assembler in 1990 and that I was looking for something similar, i.e., an integrated developer's environment for assembly. 'as' will probably do the trick, though. For those interested, Luke Adamson kindly pointed me to : http://www.intel.com/design/pro/manuals/ where you can find PDF documentation for the Pentium Pro Processor. -- thomas harte @ computer laboratory, cambridge university; tph1001@CL.cam.ac.uk; phone: +44 1223 334628; fax: 334678; http://www.CL.cam.ac.uk/users/tph1001 MIME & NeXT Mail OK.
From: Thomas Harte <tph1001@CL.cam.ac.uk> Newsgroups: comp.sys.next.programmer,comp.sys.next.software Subject: gcc optimized for Intel PPro NEXTSTEP 4.1 Date: 25 Sep 1997 09:13:02 GMT Organization: Cambridge University Message-ID: <60d9uu$703$2@lyra.csx.cam.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Does anyone know the whereabouts of gcc optimized for Intel PPro which is compatible with NEXTSTEP 4.1. I have heard that there are difficulties in compatibility between later versions of gcc (2.7.x, I think?) and NEXTSTEP 4.1 (I am currently running gcc 2.5.8!! which came bundled with the system... but it works fine...apart from shared libraries...and then there's no graphical debugger....) -- thomas harte @ computer laboratory, cambridge university; tph1001@CL.cam.ac.uk; phone: +44 1223 334628; fax: 334678; http://www.CL.cam.ac.uk/users/tph1001 MIME & NeXT Mail OK.
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 25 Sep 1997 12:50:16 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7hgb9acrb.fsf@tom.ics.ele.tue.nl> References: <60b3tp$stk$1@netty.york.ac.uk> <60bb5a$a0p@shelob.afs.com> In-reply-to: Greg_Anderson@afs.com's message of 24 Sep 1997 15:21:14 GMT In article <60bb5a$a0p@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) writes: > NSRect fr = [aView frame]; > > might crash the application with stack corruption if aView is nil. BTW, the reason for this problem is that when a method returns a struct, the compiler uses an alternate form of objc_msgSend() that passes the return value as a 3rd hidden argument on the stack (like self and _cmd). I'm sure it doesn't got to the trouble of initializing that memory, which can lead to unpredictable results if the method never gets called due to (aView == nil). Been there, done that. It will not crash the app. A method is like a function, with two extra arguments (self and _cmd) preceding the usual arguments; the latter include a pointer to enough space for a struct if one is to be returned (and only if sizeof (the struct) is larger than what is returned in registers, normally sizeof (double)). So, nothing special here and objc_msgSend can (and will) be used (checked on m68k and i386 NS3.3, and HPPA/HP-UX with gcc -fnext-runtime). Of course, the NSRect that is returned in the example will contain garbage. --Tiggr
From: brown@bibliotech.com (Robert E. Brown) Newsgroups: comp.sys.next.programmer Subject: [NSImage compositeToPoint: ...] and clipping Date: 25 Sep 1997 09:02:09 -0400 Organization: Bibliotech, Inc. Sender: brown@egil Message-ID: <87bu1hbl7y.fsf@bibliotech.com> I am using the compositeToPoint:operation: method of class NSImage with operation NSCompositeCopy but the image I'm copying does not seem to be clipped to the current clipping path. Should it be or am I just doing something wrong? If the compositeToPoint:operation: method is not supposed to obey the current clipping path, what's the easiest way to copy an image into a view and have it clipped to a path? bob
From: <webmaster@webvermont.com> Newsgroups: comp.sys.next.programmer Subject: This is a must visit site! Message-ID: <342a6ca1.0@news.together.net> Date: 25 Sep 97 13:52:33 GMT Check out this site! http://www.webvermont.com
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 25 Sep 1997 14:43:04 GMT Organization: Anderson Financial Systems Inc. Message-ID: <60dt9o$etp@shelob.afs.com> References: <x7hgb9acrb.fsf@tom.ics.ele.tue.nl> Pieter Schoenmakers writes > Someone else wrote: > > NSRect fr = [aView frame]; > > > > might crash the application with stack corruption if aView is nil. > It will not crash the app. A method is like a function, with two extra > arguments (self and _cmd) preceding the usual arguments; the latter > include a pointer to enough space for a struct if one is to be returned > (and only if sizeof (the struct) is larger than what is returned in > registers, normally sizeof (double)). So, nothing special here and > objc_msgSend can (and will) be used (checked on m68k and i386 NS3.3, and > HPPA/HP-UX with gcc -fnext-runtime). Of course, the NSRect that is > returned in the example will contain garbage. --Tiggr That's the problem. The garbage is unpredictable, and it is definitely unsafe to send things like NaN as coordinates to the windowserver. Maybe the app won't crash, but it can become so littered with PS exceptions that your only choice is to kill and restart it. Ask me how I know this... 8^) -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Forcing copies of objects over DO Date: 25 Sep 1997 15:06:12 GMT Organization: Anderson Financial Systems Inc. Message-ID: <60dul4$evu@shelob.afs.com> The WriteUp API uses DO to communicate between its API methods and a user-written "composer" app. I have a customer who wants to send an NSImage -- not the filename, but the image itself -- to WriteUp for placement in the text stream. The problem is, the image is sent as a proxy (NSDistantObject), which means there are constant references back to a real object in the "composer" app. Not only can this be slow, but if the composer app goes away, WriteUp has a hanging reference to an object that has disappeared. My question: Is there a magic incantation to force the NSImage to be copied across the wire, with no further dependency on the originator? Or is this an implementation detail built into the NSImage port coding that I can't override easily? -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
From: Joe Panico <jpanico@ml.com> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: Thu, 25 Sep 1997 10:49:21 -0400 Organization: Merrill Lynch Message-ID: <342A79F1.8C8B63E3@ml.com> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I don't believe that writing fully untyped code is a benefit at all--it's basically a poor coding proactice. Every ObjC shop I have worked in has insisted that developers strongly type wherever they can. And if you look at Next example code, you will see this pattern there too. The benefits of this approach are tremendous-- the compiler does a lot more work for you; your code is more understandable to outsiders, etc... In Java, you could always use type (Object) if you wanted to disable type checking, so in fact Java can do this as well as Objective-C. Robert A. Decker wrote: > On Mon, Sep 8, 1997 9:56 AM, Joseph Panico > <mailto:jpanico@online.disney.com> wrote: > >Here is what ObjC can do that Java cannot: > > You forgot one of the most important things: Not having to worry about > type in Obj-C. After programming in Obj-C and Newtonscript and a little > Dylan I absolutely cannot stand the static typing of Java/C++. > In fact, static typing ruins the concept of OO programming. Why should I > care what type the object is that I just got? I should just be able to tell > it to do something and leave it at that. (that being said, I usually use > static typing in Obj-C except for a few functions). > > rob > -- > <mailto: "Robert A. Decker" comrade@umich.edu> > <http://hmrl.cancer.med.umich.edu/Rob/index.ssi> > Programmer Analyst - Health Media Research Lab > University of Michigan Comprehensive Cancer Center > > "Get A Life" quote #18: "Well, you're always saying I can do anything I > want to do if I put my mind to it. This is something I can do without > putting my mind to it at all." -Chris Elliott in "The Prettiest Week of My > Life"
From: sef@kithrup.com Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <342a6ca1.0@news.together.net> Date: 25 Sep 1997 16:12:03 GMT Control: cancel <342a6ca1.0@news.together.net> Message-ID: <cancel.342a6ca1.0@news.together.net> Sender: <webmaster@webvermont.com> Spam cancelled by sef@kithrup.com
From: beigel@eecs.lehigh.edu (Richard Beigel) Newsgroups: comp.sys.next.programmer Subject: problem compiling C programs under nextstep 3.3 Date: 25 Sep 1997 17:49:37 GMT Message-ID: <60e87h$1s0u@fidoii.cc.Lehigh.EDU> Summary: please help Keywords: help I'm having a really basic problem compiling C programs. The preprocessor doesn't find the .h files in /usr/include/. Using absolute paths is an ugly hack and it doesn't work anyway, because then the next_include's fail. This is probably my fault. I think I might have deleted some files that I need but I can't remember what. I'm running nextstep 3.3 on a b/w turbo slab, if that matters.
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: Forcing copies of objects over DO Date: 25 Sep 1997 18:53:51 GMT Organization: Apple Computer, Inc. Message-ID: <60ebvv$m6g$1@news.apple.com> References: <60dul4$evu@shelob.afs.com> Gregory H. Anderson writes > My question: Is there a magic incantation to force the NSImage to be > copied across the wire, with no further dependency on the originator? > Or is this an implementation detail built into the NSImage port coding > that I can't override easily? Have you tried using the "bycopy" modifier in your method declaration? From the Objective-C docs: There are times when proxies may be unnecessarily inefficient, when it's better to send a copy of the object to the remote process so that it can interact with it directly in its own address space. To give programmers a way to indicate that this is intended, Objective-C provides a bycopy type modifier: - danceWith:(bycopy id)aClone; bycopy can also be used for return values: - (bycopy)dancer; I've never tried this with NSImage, but it certainly works for other classes. The class to be sent bycopy needs to have initWithCoder and encodeWithCoder methods, but most of the AppKit/Foundation classes do. I think that you HAVE to use setProtocolForProxy: too. -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 25 Sep 1997 19:27:19 GMT Organization: Apple Computer, Inc. Message-ID: <60edun$hfm$1@news.apple.com> References: <342A79F1.8C8B63E3@ml.com> Joe Panico <jpanico@ml.com> writes [about how strong typing is a good thing, but...] > In Java, you could always use type (Object) if you wanted to disable > type checking, so in fact Java can do this as well as Objective-C. No, that isn't the same thing at all. Using Object in Java just requires you to litter your code with casts in order to call any method that isn't defined for Object. Note that a Java cast isn't a passive type conversion like a C cast. Type checking is performed at the time of the cast. Java ALWAYS requires you to know the classs of an object to perform a method on it. (Well, not ALWAYS, exactly, but that's close enough). Objective-C NEVER requires that you know the class of an object to message it. The requirement for truly anonymous objects is pretty rare in Objective-C (in my experience), but it is something you can do Objective-C that you can't do (easily) in Java. -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: arti@lava.DOTnet (Art Isbell - remove "DOT") Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 25 Sep 1997 05:15:06 GMT Organization: LavaNet, Inc. Distribution: world Message-ID: <60cs0q$md1@mochi.lava.net> References: <60b3tp$stk$1@netty.york.ac.uk> <60bb5a$a0p@shelob.afs.com> Greg_Anderson@afs.com (Gregory H. Anderson) wrote: > BTW, the reason for this problem is that when a method returns a struct, > the compiler uses an alternate form of objc_msgSend() that passes the > return value as a 3rd hidden argument on the stack (like self and _cmd). Aha! Is this the reason why ad hoc execution of methods that return structs under gdb don't work? gdb) print [aView frame] always returns some message about an illegal cast. Has anyone figured out how to execute methods that return structs under gdb? -- Art Isbell NeXT/MIME Mail: arti@lavaDOTnet Trego Systems (for whom I don't speak) Voice/Fax: +1 808 394 0511 OPENSTEP/NT Voice Mail: +1 808 394 0495 managed care solutions US Mail: Honolulu, HI 96825-2638
From: bettis@inetnebr.com (Mr. Jeremy Bettis) Newsgroups: comp.sys.next.programmer Subject: Rapsody Release Sept 24 ??? Date: 25 Sep 1997 15:37:06 -0500 Organization: Internet Nebraska Message-ID: <60ei1i$t99$1@falcon.inetnebr.com> NNTP-Posting-User: bettis Was I just dreaming, or did I hear that Developer Release 1 was to be released Sept 24? I look at www.apple.com and www.next.com, and I saw no mention. I guess I was expecting a press release or a parade or fireworks or something. I suppose I'll just wait for the mail to come...
From: "CALL-NOW !!!" email97@email97now1.com Newsgroups: comp.sys.next.programmer Subject: >>> F R E E V A C A T I O N S <<< Date: Thu, 25 Sep 1997 19:41:19 -0700 Organization: CDEF Message-ID: <250997194119@email97now1.com> ¤ ¤ ¤ F R E E V A C A T I O N S !!! ¤ ¤ ¤ 3 Day / 2 Night Vacation for up to a family of 5 http://www.vacationpromotions.com =<=<<=>=<<<<<>>>=<><
From: Steve Dekorte <dekorte@slip.net> Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? Date: 25 Sep 1997 23:57:48 GMT Organization: Slip.Net (http://www.slip.net) Message-ID: <60etps$bqd$2@owl.slip.net> References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <342A79F1.8C8B63E3@ml.com> Joe Panico <jpanico@ml.com> wrote: > In Java, you could always use type (Object) if you wanted to disable type > checking, so in fact Java can do this as well as Objective-C. Oh yeah? Can I send any message to object without the compiler choking? -- Steve Dekorte - OpenStep consultant - San Francisco
From: spamcancel@wupper.com Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <250997194119@email97now1.com> Control: cancel <250997194119@email97now1.com> Date: 26 Sep 1997 00:18:57 GMT Message-ID: <cancel.250997194119@email97now1.com> Sender: "CALL-NOW !!!" email97@email97now1.com Excessive Multi-Posted spam article exceeding a BI of 20 cancelled by spamcancel@wupper.com. From was: "CALL-NOW !!!" email97@email97now1.com Subject was: >>> F R E E V A C A T I O N S <<< NNTP-Posting-Host was: dd05-125.dub.compuserve.com
From: mpaque.spa-am@nospam.wco.com (Mike Paquette) Newsgroups: comp.sys.next.programmer Subject: Re: Implementation behind NSLocks?? Date: Wed, 24 Sep 1997 06:24:57 GMT Organization: Electronics Service Unit No. 16 Message-ID: <3428af07.4181502@news.wco.com> References: <606mic$pqf$1@xenon.cube.de> <607t7a$opu$1@news.tudelft.nl> On 23 Sep 1997 08:04:58 GMT, guyt@is.twi.tudelft.nl (A. Guyt) wrote: >This is really, really strange... It is simplly nuts to forbid the >creation of locks after the creation of any thread !! One cannot write >serious threaded apps with this restriction. Any object created at run >time on the fly that needs to be thread safe needs its own lock !! Nor >Mach nor Windows NT has this restriction of its own. > >It is even stranger that Apple's own documentation contradicts this >restriction. Look at the NSlocking protocol documentation. The example >included shows an object that calls a method creating a lock after the >objects detect the multithreaded state ! The documentation is just trying to establish guidelines to keep unexperienced programmers from stepping on their... code. Some tech writer seems to have gotten carried away. Imagine the case of a multithreaded app in which one thread allocates a lock, and another thread tries to reference the lock before it is initialized. Imagine the case where the code does something of the form: if ( lock == nil ) [lock = [[NSLock alloc] init] lock]; If two threads hit this code with just the right timing, the app could leak a lock, and two threads would run through the critical section at the same time. The experienced programmer will know how to avoid such problems... Mike Paquette mpaque AT wco.com (Damn junk-mailers!)
From: eric@bacchus (Eric Bloom) Newsgroups: comp.sys.next.programmer Subject: Compiler version under OS 4.2 ? Date: 26 Sep 1997 03:19:45 GMT Organization: MindSpring Enterprises Message-ID: <60f9kh$3dp@camel12.mindspring.com> Has the compiler been updated from OpenStep/MACH dev. release 4.1 to OpenStep 4.2/MACH. The version under OS 4.1/MACH is 2.5.8. Also, have they made it easier to compile both Objective C and C++ within Project Builder without making customizations. Eric Bloom Bacchus, Inc. eric@bacchus.com
From: arti@lava.DOTnet (Art Isbell - remove "DOT") Newsgroups: comp.sys.next.programmer Subject: Crash trap testers needed Date: 26 Sep 1997 04:33:12 GMT Organization: LavaNet, Inc. Distribution: world Message-ID: <60fdu9$fbv@mochi.lava.net> I need some assistance in testing the latest incarnation of Julie Zelenski's ObjectError (described below). I plan to post it at archive sites once I'm fairly certain that it works well. I've posted source for a beta version at ftp.lava.net/users/arti that's available via anonymous FTP. I would appreciate those of you interested in using a crash trap with your projects testing OEObject and providing feedback. OEObject An OPENSTEP Crash Trap Art Isbell arti@lava.net 25 Sep 1997 Description OEObject is a free NSObject class poser that is designed to trap several types of application errors that would normally lead to an abnormal termination. These errors include various fatal signals, messages sent to freed objects, and messages that aren't recognized by the receiver. OEObject can be used to allow the user to gracefully exit the app, possibly saving changes, instead of the app just disappearing or the dreaded Dr. Watson (Windows) rearing its ugly head as the app disappears. OEObject also logs a stack backtrace to help identify a programming error even when a process isn't run under gdb. OEObject was originally a NEXTSTEP class called ObjectError which morphed into HKCrashTrap, at least one commercial crash catcher, and possibly other versions, but it has now been ported to OPENSTEP as OEObject. Usage Simply add OEObject to an app or tool project or to the project of a framework that is loaded by apps and tools. Then send OEObject a setup message early in the process' execution. -- Art Isbell NeXT/MIME Mail: arti@lavaDOTnet Trego Systems (for whom I don't speak) Voice/Fax: +1 808 394 0511 OPENSTEP/NT Voice Mail: +1 808 394 0495 managed care solutions US Mail: Honolulu, HI 96825-2638
From: "Lee Byeong-ho" <bhlee@cnt.co.kr> Newsgroups: comp.sys.next.programmer Subject: I can't start apache server. Date: 26 Sep 1997 05:30:34 GMT Organization: HanQ.Net Message-ID: <01bcca3d$3ab72c30$68a67fd2@dooly> Hi, We have a fire wall and I configured OmniWeb Proxy Preference( Proxy Server URL : http://myproxy server: 8080). I gave command '/NextLibrary/WebServer/httpd -d /NextLibrary/WebServer > /dev/console', but I got error message 'httpd: could not bind to port 8080, bind: Address already in use' I can't start server. And I can't find domain name.( not ip address, I wrote nameserver information in /etc/resolv.conf) Please give me some tips. Thanks in advance... /BYEONG HO LEE
From: pekeler@luck.shnet.org (Christian Pekeler) Newsgroups: comp.sys.next.programmer Subject: Re: Compiler version under OS 4.2 ? Date: 26 Sep 1997 08:25:38 GMT Organization: private Message-ID: <60fri2$is@luck.shnet.org> References: <60f9kh$3dp@camel12.mindspring.com> eric@bacchus (Eric Bloom) wrote: >Has the compiler been updated from OpenStep/MACH dev. release 4.1 to >OpenStep 4.2/MACH. The version under OS 4.1/MACH is 2.5.8. luck> cc -v Reading specs from /lib/i386/specs NeXT Software, Inc. version cc-744.13, gcc version 2.7.2.1 >Also, have they made it easier to compile both Objective C and C++ within >Project Builder without making customizations. Don't know what you are talking about, and the ProjectBuilder release notes don't contain something interesting about c++. Maybe the following quote from the Compiler release notes is of interest to you: >> The header files on OPENSTEP and PDO 4.2 have been sanitized and no longer contain uses of C++ keywords as parameter names, struct field names, or function names. This should make C++ usage easier on OPENSTEP and more similar to other C++ development environments. << Christian
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: Setting nil objects as values for UI elements...wierd. Date: 26 Sep 1997 11:10:56 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7wwk4xwwv.fsf@tom.ics.ele.tue.nl> References: <60b3tp$stk$1@netty.york.ac.uk> <60bb5a$a0p@shelob.afs.com> <60cs0q$md1@mochi.lava.net> In-reply-to: arti@lava.DOTnet's message of 25 Sep 1997 05:15:06 GMT In article <60cs0q$md1@mochi.lava.net> arti@lava.DOTnet (Art Isbell - remove "DOT") writes: Aha! Is this the reason why ad hoc execution of methods that return structs under gdb don't work? It could indeed be a gdb mixup concerning the location, in the call frame, of a third implicit argument and two other slightly different implicit arguments. gdb) print [aView frame] always returns some message about an illegal cast. Has anyone figured out how to execute methods that return structs under gdb? Eh, what about: (gdb) print (NSRect *) malloc (sizeof (NSRect)) $1 = ... (gdb) call objc_msgSend (aView, sel_getUid ("frame"), $) (gdb) print *$1 (Untested, but the basic idea should work.) --Tiggr
From: guyt@is.twi.tudelft.nl (A. Guyt) Newsgroups: comp.sys.next.programmer Subject: How does Rhapsody deal with Java vs. Objective-C? Date: 26 Sep 1997 10:07:22 GMT Organization: Delft University of Technology Message-ID: <60g1gq$7fd$1@news.tudelft.nl> References: <60edun$hfm$1@news.apple.com> L.S., We've seen some nice discussions about Java and Obj-C. I think we know the languages differences by now. What I'd like to now is: when Apple provides the Java language to program and use all the Yellow Box api's, how is dealt with the dynamic nature of the Yellow Box ? Does the use of a static language like Java not generate problems in conjunction with these ? for example: - how will distributed objects work when there is no forwarding in Java ? Hope to get useful answers, Abraham Guyt. _____________________________________________________________________ Abraham Guyt P.O.Box 356 Department of Information Systems 2600 AJ Delft Faculty Information Technology & Systems The Netherlands Delft University of Technology tel: +31 15 278 5969 E-mail: guyt@is.twi.tudelft.nl NeXT-mail welcome
From: tiggr@ics.ele.tue.nl (Pieter Schoenmakers) Newsgroups: comp.sys.next.programmer Subject: Re: How does Rhapsody deal with Java vs. Objective-C? Date: 26 Sep 1997 12:55:51 +0200 Organization: Eindhoven University of Technology Sender: tiggr@tom.ics.ele.tue.nl Message-ID: <x7vhzoxs20.fsf@tom.ics.ele.tue.nl> References: <60edun$hfm$1@news.apple.com> <60g1gq$7fd$1@news.tudelft.nl> In-reply-to: guyt@is.twi.tudelft.nl's message of 26 Sep 1997 10:07:22 GMT In article <60g1gq$7fd$1@news.tudelft.nl> guyt@is.twi.tudelft.nl (A. Guyt) writes: Does the use of a static language like Java not generate problems in conjunction with these ? The dynamic functionality needed by Java is a subset of that needed by Objective-C. There is no semantic mismatch, so there should be no problem (not that I am in a position to judge; call this an educated judgement). - how will distributed objects work when there is no forwarding in Java ? Why would you need forwarding in Java code, if the runtime used were Objective-C, and the proxy objects were Objective-C objects? Hope to get useful answers, Don't fear the dark if dusk is still to come. --Tiggr
From: guyt@is.twi.tudelft.nl (A. Guyt) Newsgroups: comp.sys.next.programmer Subject: Re: How does Rhapsody deal with Java vs. Objective-C? Date: 26 Sep 1997 11:38:43 GMT Organization: Delft University of Technology Message-ID: <60g6s3$9i3$1@news.tudelft.nl> References: <x7vhzoxs20.fsf@tom.ics.ele.tue.nl> Pieter Schoenmakers writes > In article <60g1gq$7fd$1@news.tudelft.nl> guyt@is.twi.tudelft.nl (A. Guyt) writes: > > Does the use of a static language like Java not generate problems in > conjunction with these ? > > The dynamic functionality needed by Java is a subset of that needed by > Objective-C. There is no semantic mismatch, so there should be no problem > (not that I am in a position to judge; call this an educated judgement). > > - how will distributed objects work when there is no forwarding in Java ? > > Why would you need forwarding in Java code, if the runtime used were > Objective-C, and the proxy objects were Objective-C objects? > > Hope to get useful answers, > > Don't fear the dark if dusk is still to come. --Tiggr Well consider this: I want to send a message to a proxy wihtin a protocol. In Obj-C you can do this right away using the NSProxy class. In Java this is impossible since you can't assing a protocol dynamically at run time to the NSProxy class (java version). In java you would need to subclass NSProxy to a class that implements the protocol(s) you want to use. But this move would cause other major changes to the NXConnection class.... and so on... Abraham Guyt. _____________________________________________________________________ Abraham Guyt P.O.Box 356 Department of Information Systems 2600 AJ Delft Faculty Information Technology & Systems The Netherlands Delft University of Technology tel: +31 15 278 5969 E-mail: guyt@is.twi.tudelft.nl NeXT-mail welcome
From: jwdb@fygir.nl (Jan-Willem de Bruijn) Newsgroups: comp.sys.next.programmer Subject: Makefiles 4.0 -> 4.2 Date: 26 Sep 1997 13:27:36 GMT Organization: NLnet Message-ID: <60gd88$4bf$1@news.Leiden.NL.net> Is it me, or is upgrading your Makefiles from 4.0 to 4.2 (on Mach) not a trivial matter? ProjectBuilder dutifully alerts you that it is going to do so for you, but it seems to do an incomplete job. I have the same problem as I had when I built my first project on 4.1 for NT, after having ported it from 3.3 Mach, namely: In a subproject, pswrap tries to output its files into a not yet existing subdirectory of derived_src. In the end I patched the implicitrules.make file in /NextDeveloper/Makefiles/pb_makefiles with an extra command line @$(MKDIRS) $(SFILE_DIR) in the rules .psw.c and .psw.h. This does the trick but if anybody knows how it should work, I'll be glad to hear from them. Chao, Jan-Willem -- Jan-Willem de Bruijn - F Y G I R logistic information systems http://www.fygir.com/
From: marcel@system.de Newsgroups: comp.sys.next.programmer Subject: Re: How does Rhapsody deal with Java vs. Objective-C? Date: 26 Sep 1997 18:42:49 GMT Organization: Technical University Berlin, Germany Distribution: world Message-ID: <60gvn9$shn$1@brachio.zrz.TU-Berlin.DE> References: <60g6s3$9i3$1@news.tudelft.nl> In article <60g6s3$9i3$1@news.tudelft.nl> guyt@is.twi.tudelft.nl (A. Guyt) writes: > > Well consider this: I want to send a message to a proxy wihtin a protocol. > In Obj-C you can do this right away using the NSProxy class. In Java this > is impossible since you can't assing a protocol dynamically at run time to > the NSProxy class (java version). In java you would need to subclass > NSProxy to a class that implements the protocol(s) you want to use. But > this move would cause other major changes to the NXConnection class.... > and so on... I am not quite sure if I got your meaning correctly, but it seems to be that you are talking about the _implementation_ of DO (proxies etc.) That one is easy: the implementation is already there (in Obj-C), Java simply uses it. Implementing/extending DO in Jave would probably be somewhat more challenging... Marcel
From: email@end.of.post (Raymond Lutz) Newsgroups: comp.sys.next.programmer Subject: waiting a NSNotification before returning? Date: 26 Sep 1997 21:29:10 GMT Organization: VTL Message-ID: <60h9f6$6sa$1@wagner.videotron.net> Bonjour chere communaute Openstep I have a method that will submit some bytes to a MiscSerialPort instance to send through the serial port and I don't want this method to return until an ACK byte is received on the port (or not, after some timeout delay). The problem is that arrival of data on the serial port is signaled through a NSNotification (delegation would not change anything anyway, the problem is asynchronousness). How can a method not return right away but still permits NSNotification reception? Merci pour toute suggestion! And thanks to Don for this great MiscKit! And W. Eric Norum for the MiscSerialPort code ! Ray -- Raymond Lutz - lutzray-at-9bit.qc.ca - www.9bit.qc.ca/~lutzray - "Les 400 plus fortunes individus de la planete possedent autant que 2.3 MILLIARDS des plus pauvres reunis"
From: dwy@ace.net (David Young) Newsgroups: comp.sys.next.programmer Subject: Re: waiting a NSNotification before returning? Date: 26 Sep 1997 22:18:09 GMT Organization: 21st Century Software, New York City Sender: daver@138.202.212.111 Message-ID: <60hcb1$kr5$1@darla.visi.com> References: <60h9f6$6sa$1@wagner.videotron.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Date: 26 Sep 1997 17:18:09 CDT Cc: email@end.of.post In <60h9f6$6sa$1@wagner.videotron.net> Raymond Lutz wrote: > I have a method that will submit some bytes to a MiscSerialPort > instance to send through the serial port and I don't want this > method to return until an ACK byte is received on the port (or not, > after some timeout delay). The problem is that arrival of data on > the serial port is signaled through a NSNotification (delegation > would not change anything anyway, the problem is asynchronousness). > > How can a method not return right away but still permits > NSNotification reception? You may not be able to do this with a notification; however: @implementation MiscSerialPort(FileHandleAccess) - (NSFileHandle *)fileHandle { return fileHandle; } @end Then use [[serialPort fileHandle] readDataOfLength:1] to confirm the ACK byte at the end of your write method. Hope this helps, Dave -- :: d a v i d y o u n g ::::: smtp dwy@ace.net http www.ace.net ::
From: Chris Roehrig <croehrig@House.ORG> Newsgroups: comp.sys.next.programmer,comp.unix.internals Subject: How do I get a (struct proc *) for the current user process (in NEXTSTEP 3.3)? Date: 26 Sep 1997 23:03:21 GMT Organization: Computer Science, Univerity of B.C., Vancouver, B.C., Canada Message-ID: <60hevp$7va$1@nnrp.cs.ubc.ca> I'm writing a device driver for NEXTSTEP 3.3 (Mach 2.5-based) and I need to get the struct proc * for the current user process from within an ioctl system call in order to send that process a signal later on (via psignal). (I know Mach prefers to use messages, but that's the user API I'm stuck with.) How do I get the struct proc* for the user process that performed the ioctl? It doesn't seem to be passed as a parameter to my device_ioctl entry point as it does in other BSD unices. Where do I learn more about these NeXT Mach internals? Thanks, Chris -- Chris Roehrig croehrig@House.ORG Neuroscience and Computer Science at University of British Columbia, Vancouver http://www.House.ORG/chris http://www.sns.cs.ubc.ca/chris
From: george@ee.ualberta.ca (Jason'Curious'George) Newsgroups: comp.sys.next.misc,comp.sys.next.programmer Subject: 4.2 availability Date: 23 Sep 1997 16:25:13 GMT Organization: University of Alberta Electrical Engineering Department Message-ID: <608qh9$m6c$2@pulp.ucs.ualberta.ca> Can someone confirm the availability of Openstep 4.2? I can't seem to locate much information about it from the NeXT homepage and the net resources I've found are vague at best. Information that I've sucked from colleagues is pretty contradictory. I'd like to buy the latest release of the full development environment on the academic discount. I'd prefer the full-on NeXT environment, but the NT environment will do if need be. The platform is x86. Can anyone offer any tangible information? Thanks in advance. --Jason j.b.george@ieee.org george@ee.ualberta.ca
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: 23 Sep 1997 21:16:09 GMT Organization: Anderson Financial Systems Inc. Message-ID: <609bip$66k@shelob.afs.com> References: <6096im$dfi$1@netty.york.ac.uk> -bat. writes > mmalcolm crawford <malcolm@plsys.co.uk> writes: > > If you are running OPENSTEP/4.x Mach go to the Termnial and do: > > > > defaults write NSGlobalDomain NSInterfaceStyle Windows95 > > ..and then spend 3/4 of an hour trying to fine anything which has > changed in the slightest ! The only noticeable thing I could see was > that the scrollbars on Interface Builders window has changed. Apart > from that everything was identical. Am I missing something ? I expected > to see OpenStep applications looking more "windows 95"ish somehow... You need to logout and back in to see the whole effect, because running apps do not respond to this change immediately. Trust me, it's more than just the scrollbars (which lose the dimple, move to the right, and adopt a new "elevator click" behavior). All widgets -- checkboxes, radios, and so on -- have an entirely different look. The colors inside scrolling matrices are "reversed" too. -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
From: Paul Kapcio <paul.kapcio@lmco.com> Newsgroups: sci.space.tech,comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE F***ING BOARD!! Date: Thu, 25 Sep 1997 15:01:52 -0400 Organization: Lockheed Martin Federal Systems - Manassas Message-ID: <342AB520.54C7@lmco.com> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.ps i.net> <01bcba15$447c2f00$48e04382@broennic.online.no> <acurylo-0509971010020001@p2-29.van.tvs.net> <5ur2uv$r3s@crcnis3.unl.edu> <5uurkb$d74@gap.cco.caltech.edu> <5v1sq9$dfb$1@news.jumpnet.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Delivery-Date: Thu, 25 Sep 1997 12:17:46 -0700 Return-Path: usenet@postoffice.man.fs.lmco.com for <gherbert@crl.com>; Thu, 25 Sep 1997 11:59:55 -0700 (PDT) env-from (usenet@postoffice.man.fs.lmco.com) id QQdioi20432; Thu, 25 Sep 1997 15:05:07 -0400 (EDT) id sma014188; Thu Sep 25 15:04:29 1997 To: sci-space-tech@uunet.uu.net Tim Olson wrote: > > In article <5uurkb$d74@gap.cco.caltech.edu> > madler@alumni.caltech.edu (Mark Adler) writes: > > > In article <5ur2uv$r3s@crcnis3.unl.edu>, > > Josh Hesse <00093182@bigred.unl.edu> wrote: > > >doesn't the Mars Rover use a version of the Intel 8085? > > > > Yep. A rad-hard 80C85. Pathfinder doesn't have a PPC either. > > Well, you are correct in the strictest sense of "PPC". But Pathfinder > *does* have a rad-hard RS/6000 on board: > > [From IBM's web pages]: > ---- <snip> > -- Tim Olson The MPF computer contains the Lockheed Martin Federal Systems - Manasss RAD-6000 CPU. LM designed the computer board and manufactured the RAD-6000 CPU. The CPU design is a transfer of the IBM RSC (RISC single chip) processor that was the first in the lineage of PowerPC chips. The RSC begat the PowerPC 601 (RSC + Motorola 88000 bus + misc.). It has no connection to the MIPS Rx000 processors as was stated in several earlier posts. When the RAD-6000 was first started we were part of IBM; w hen we delivered the RAD-6000 we were part of Loral and when it landed we were part of LM ;-) This has caused part of the confusion as to who did what. Paul Kapcio Applications Engineering for Space Systems Lockheed Martin Federal Systems - Manassas
From: email@end.of.post (Raymond Lutz) Newsgroups: comp.sys.next.programmer Subject: Re: waiting a NSNotification before returning? Date: 27 Sep 1997 04:38:35 GMT Organization: VTL Message-ID: <60i2kb$ifm$1@wagner.videotron.net> References: <60h9f6$6sa$1@wagner.videotron.net> <60hcb1$kr5$1@darla.visi.com> In-Reply-To: <60hcb1$kr5$1@darla.visi.com> On 09/26/97, David Young wrote: >In <60h9f6$6sa$1@wagner.videotron.net> Raymond Lutz wrote: >> I have a method that will submit some bytes to a MiscSerialPort >> instance to send through the serial port and I don't want this >> method to return until an ACK byte is received on the port (or >> not, after some timeout delay). The problem is that arrival of >> data on the serial port is signaled through a NSNotification >> (delegation would not change anything anyway, the problem is >> asynchronousness). >> >> How can a method not return right away but still permits >> NSNotification reception? > >You may not be able to do this with a notification; however: > >@implementation MiscSerialPort(FileHandleAccess) - (NSFileHandle >*)fileHandle { > return fileHandle; >} @end > >Then use [[serialPort fileHandle] readDataOfLength:1] to confirm >the ACK byte at the end of your write method. > >Hope this helps, Merci! So I guess from what you're suggesting (it's not stated in the docs) that readDataOfLength: will block until some data arrives, like availableData. But How can I set a timeout on that? 8^| A defective device on the port shouldn't block my program... Hmm.. maybe should I return from [myMiscSerialPort writeData:data], set a NSTimer and wait for one of those: either ACK reception or timer completion. The latter arriving first would determine a timeout condition. Ray -- Raymond Lutz - lutzray @foobar@ 9bit.qc.ca - www.9bit.qc.ca/~lutzray - "Les 400 plus fortunes individus de la planete possedent autant que 2.3 MILLIARDS des plus pauvres reunis"
From: jonathan@quern.demon.co.uk (Jonathan Sanderson) Newsgroups: comp.sys.next.programmer,comp.sys.next.advocacy,comp.sys.mac.programmer,comp.sys.mac.advocacy Subject: Re: Openstep 4.2 / Prelude to Rhapsody "Easter Egg" Date: Sat, 27 Sep 1997 15:29:48 +0100 Message-ID: <199709271529481164400@quern.demon.co.uk> References: <5vds3m$di4$1@ironhorse.plsys.co.uk> <6096im$dfi$1@netty.york.ac.uk> <60b3v1$4oe$1@ironhorse.plsys.co.uk> mmalcolm crawford <malcolm@plsys.co.uk> wrote: > On 09/23/97, -bat. wrote: > >everything was identical. Am I missing something ? I expected to see > >OpenStep applications looking more "windows 95"ish somehow... > > > I think maybe you need to restart relevant (4.2) applications... > Try logging out then in again. > > If this fails, maybe a visit to SpecSaver...? :-) Are you suggesting that looking at the Win95 GUI is like looking through blurred, poorly-ground lenses, with horrid chromatic aberrations and an inaccurate correction for astigmatism? [squints] It's not *that* ugly. ;-) (YMMV with SpecSaver. I hope) -- Jonathan Sanderson <jonathan@quern.demon.co.uk> Specialist Researcher, science TV <http://www.quern.demon.co.uk> If I had more time, I would have written you a shorter letter (Pascal)
From: mtrombin@ix.netcom.com (Mark Trombino) Newsgroups: comp.sys.next.programmer Subject: Flushing Windows Date: 27 Sep 1997 18:18:00 GMT Organization: Egghead Billy, Inc. Message-ID: <60jiko$d1k@dfw-ixnews12.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii I'm having a problem with getting a window to update properly before a system call that blocks. Even though setStringValue: is called before the system call, and even if I send the message flushWindow to the window itself before the system call, the window doesn't get updated until after the system call finishes. I've tried to use DPSFlush() but that seems to be missing in my installation (Openstep 4.1). I've also tried using a non buffered window, with the same results. - (void)setPPPActive:(BOOL)flag { if(flag) { // Send queued mail if user wants if([[self objectForKey:LOGON_SENDMAILKEY] intValue]) { [messagePanel displayMessage:@"Sending queued mail..."]; system("/usr/lib/sendmail -q"); } } [super setPPPActive:flag]; } - (void)displayMessage:(NSString *)message { if(panel) { [messageText setStringValue:message]; [panel flushWindow]; } } Is there a way I can get that window updated before the system call? Thanks! -- Mark Trombino mtrombin@ix.netcom.com (NEXTMail, MIME Mail okay)
From: dwy@ace.net (David Young) Newsgroups: comp.sys.next.programmer Subject: Re: Flushing Windows Date: 27 Sep 1997 22:17:04 GMT Organization: 21st Century Software, New York City Sender: daver@138.202.212.111 Message-ID: <60k0l0$kcf$1@darla.visi.com> References: <60jiko$d1k@dfw-ixnews12.ix.netcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Date: 27 Sep 1997 17:17:04 CDT Cc: mtrombin@ix.netcom.com In <60jiko$d1k@dfw-ixnews12.ix.netcom.com> Mark Trombino wrote: > I'm having a problem with getting a window to update properly before a system > call that blocks. Even though setStringValue: is called before the system > call, and even if I send the message flushWindow to the window itself before > the system call, the window doesn't get updated until after the system call > finishes. I've tried to use DPSFlush() but that seems to be missing in my > installation (Openstep 4.1). I've also tried using a non buffered window, > with the same results. [snip] > Is there a way I can get that window updated before the system call? Try [panel update]. BTW, the equivalent of DPSFlush is now [[NSDPSContext currentContext] flush]. -- :: d a v i d y o u n g ::::: smtp dwy@ace.net http www.ace.net ::
From: sanguish@digifix.com Newsgroups: comp.sys.next.advocacy,comp.sys.next.announce,comp.sys.next.hardware,comp.sys.next.software,comp.sys.next.misc,comp.sys.next.sysadmin,comp.sys.next.bugs,comp.sys.next.programmer Subject: NEXTSTEP/OpenStep Resources on the Net Supersedes: <12335874814429@digifix.com> Date: 28 Sep 1997 03:48:51 GMT Organization: Digital Fix Development Message-ID: <27562875419231@digifix.com> Topics include: Major OpenStep/NEXTSTEP World Wide Web Sites OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups Major OpenStep/NEXTSTEP FTP sites NeXTanswers Major OpenStep/NEXTSTEP World Wide Web Sites ============================================ The following sites are a sample of the OpenStep related WWW sites available. A comprehensive list is available on Stepwise. Stepwise OpenStep/Rhapsody Information Server http://www.stepwise.com Stepwise has been serving the OpenStep/NEXTSTEP community since March 1993. Some of the many resources on the site include: OpenStep Third Party Software guide, Developer Directory, Mailing List information, extensive listing of FTP and WWW sites related to OpenStep and NEXTSTEP, OpenStep related Frequently Asked Questions. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. NeXT Software Archives @ Peak.org http://www.peak.org/next http://www.peak.org/openstep PEAK is the premier NeXTStep/OpenStep FTP site in North America. This is the World Wide Web interace to the FTP site. Apple Enterprise Software Group (formerly NeXT Computer, Inc.) http://www.next.com Here is where you'll find the NeXTanswers archive, with information on OpenStep installation, drivers and software patches. Apple Computer's 'Prelude to Rhapsody' Self Support Site http://devworld.apple.com/dev/prelude.html This site has been constructed to help you help yourself to learn as much as possible about the foundation for Rhapsody, today's OPENSTEP. The site provides an informal collection of pointers, references, and starting points for developers who are using the Prelude to Rhapsody bundle, distributed at this year's Worldwide Developer Conference. OpenStep/NEXTSTEP/Rhapsody Related Usenet Newsgroups ==================================================== COMP.SYS.NEXT.ADVOCACY This is the "why NEXTSTEP is better (or worse) than anything else in the known universe" forum. It was created specifically to divert lengthy flame wars from .misc. COMP.SYS.NEXT.ANNOUNCE Announcements of general interest to the NeXT community (new products, FTP submissions, user group meetings, commercial announcements etc.) This is a moderated newsgroup, meaning that you can't post to it directly. Submissions should be e-mailed to next-announce@digifix.com where the moderator (Scott Anguish) will screen them for suitability. Messages posted to announce should NOT be posted or crossposted to any other comp.sys.next groups. COMP.SYS.NEXT.BUGS A place to report verifiable bugs in NeXT-supplied software. Material e-mailed to Bug_NeXT@NeXT.COM is not published, so this is a place for the net community find out about problems when they're discovered. This newsgroup has a very poor signal/noise ratio--all too often bozos post stuff here that really belongs someplace else. It rarely makes sense to crosspost between this and other c.s.n.* newsgroups, but individual reports may be germane to certain non-NeXT-specific groups as well. COMP.SYS.NEXT.HARDWARE Discussions about NeXT-label hardware and compatible peripherals, and non-NeXT-produced hardware (e.g. Intel) that is compatible with NEXTSTEP. In most cases, questions about Intel hardware are better asked in comp.sys.ibm.pc.hardware. Questions about SCSI devices belong in comp.periphs.scsi. This isn't the place to buy or sell used NeXTs--that's what .marketplace is for. COMP.SYS.NEXT.MARKETPLACE NeXT stuff for sale/wanted. Material posted here must not be crossposted to any other c.s.n.* newsgroup, but may be crossposted to misc.forsale.computers.workstation or appropriate regional newsgroups. COMP.SYS.NEXT.MISC For stuff that doesn't fit anywhere else. Anything you post here by definition doesn't belong anywhere else in c.s.n.*--i.e. no crossposting!!! COMP.SYS.NEXT.PROGRAMMER Questions and discussions of interest to NEXTSTEP programmers. This is primarily a forum for advanced technical material. Generic UNIX questions belong elsewhere (comp.unix.questions), although specific questions about NeXT's implementation or porting issues are appropriate here. Note that there are several other more "horizontal" newsgroups (comp.lang.objective-c, comp.lang.postscript, comp.os.mach, comp.protocols.tcp-ip, etc.) that may also be of interest. COMP.SYS.NEXT.SOFTWARE This is a place to talk about [third party] software products that run on NEXTSTEP systems. COMP.SYS.NEXT.SYSADMIN Stuff relating to NeXT system administration issues; in rare cases this will spill over into .programmer or .software. ** RELATED NEWSGROUPS ** COMP.SOFT-SYS.NEXTSTEP Like comp.sys.next.software and comp.sys.next.misc combined. Exists because NeXT is a software-only company now, and comp.soft-sys is for discussion of software systems with scope similar to NEXTSTEP. COMP.LANG.OBJECTIVE-C Technical talk about the Objective-C language. Implemetations discussed include NeXT, Gnu, Stepstone, etc. COMP.OBJECT Technical talk about OOP in general. Lots of C++ discussion, but NeXT and Objective-C get quite a bit of attention. At times gets almost philosophical about objects, but then again OOP allows one to be a programmer/philosopher. (The original comp.sys.next no longer exists--do not attempt to post to it.) Exception to the crossposting restrictions: announcements of usenet RFDs or CFVs, when made by the news.announce.newgroups moderator, may be simultaneously crossposted to all c.s.n.* newsgroups. Getting the Newsgroups without getting News =========================================== Thanks to Michael Ross at antigone.com, the main NEXTSTEP groups are now available as a mailing list digest as well. next-nextstep next-advocacy next-announce next-bugs next-hardware next-marketplace next-misc next-programmer next-software next-sysadmin object lang-objective-c (For a full description, send mail to listserv@antigone.com). The subscription syntax is essentially the same as Majordomo's. To subscribe, send a message to *-request@lists.best.com saying: subscribe where * is the name of the list e.g. next-programmer-request@lists.best.com Major OpenStep/NEXTSTEP FTP sites ================================= ftp://ftp.next.peak.org The main site for North American submissions formerly ftp.cs.orst.edu ftp://ftp.peanuts.org: (Peanuts) Located in Germany. Comprehensive archive site. Very well maintained. ftp://ftp.nluug.nl/pub/comp/next NeGeN/NiNe (NEXTSTEP Gebruikers Nederland/NeXTSTEP in the Netherlands) ftp://cube.sm.dsi.unimi.it (Italian NEXTSTEP User Group) ftp://ftp.nmr.embl-heidelberg.de/pub/next eduStep ftp://ftp.next.com: See below ftp.next.com and NextAnswers@next.com ===================================== [from the document ftp://ftp.next.com/pub/NeXTanswers/1000_Help] Welcome to the NeXTanswers information retrieval system! This system allows you to request online technical documents, drivers, and other software, which are then sent to you automatically. You can request documents by fax or Internet electronic mail, read them on the world-wide web, transfer them by anonymous ftp, or download them from the BBS. NeXTanswers is an automated retrieval system. Requests sent to it are answered electronically, and are not read or handled by a human being. NeXTanswers does not answer your questions or forward your requests. USING NEXTANSWERS BY E-MAIL To use NeXTanswers by Internet e-mail, send requests to nextanswers@next.com. Files are sent as NeXTmail attachments by default; you can request they be sent as ASCII text files instead. To request a file, include that file's ID number in the Subject line or the body of the message. You can request several files in a single message. You can also include commands in the Subject line or the body of the message. These commands affect the way that files you request are sent: ASCII causes the requested files to be sent as ASCII text SPLIT splits large files into 95KB chunks, using the MIME Message/Partial specification REPLY-TO address sets the e-mail address NeXTanswers uses These commands return information about the NeXTanswers system: HELP returns this help file INDEX returns the list of all available files INDEX BY DATE returns the list of files, sorted newest to oldest SEARCH keywords lists all files that contain all the keywords you list (ignoring capitalization) For example, a message with the following Subject line requests three files: Subject: 2101 2234 1109 A message with this body requests the same three files be sent as ASCII text files: 2101 2234 1109 ascii This message requests two lists of files, one for each search: Subject: SEARCH Dell SCSI SEARCH NetInfo domain NeXTanswers will reply to the address in your From: line. To use a different address either set your Reply-To: line, or use the NeXTanswers command REPLY-TO If you have any problem with the system or suggestions for improvement, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY FAX To use NeXTanswers by fax, call (415) 780-3990 from a touch-tone phone and follow the instructions. You'll be asked for your fax number, a number to identify your fax (like your phone extension or office number), and the ID numbers of the files you want. You can also request a list of available files. When you finish entering the file numbers, end the call and the files will be faxed to you. If you have problems using this fax system, please call Technical Support at 1-800-848-6398. You cannot use the fax system outside the U.S & Canada. USING NEXTANSWERS VIA THE WORLD-WIDE WEB To use NeXTanswers via the Internet World-Wide Web connect to NeXT's web server at URL http://www.next.com. USING NEXTANSWERS BY ANONYMOUS FTP To use NeXTanswers by Internet anonymous FTP, connect to FTP.NEXT.COM and read the help file pub/NeXTanswers/README. If you have problems using this, please send mail to nextanswers-request@next.com. USING NEXTANSWERS BY MODEM To use NeXTanswers via modem call the NeXTanswers BBS at (415) 780-2965. Log in as the user "guest", and enter the Files section. From there you can download NeXTanswers documents. FOR MORE HELP... If you need technical support for NEXTSTEP beyond the information available from NeXTanswers, call the Support Hotline at 1-800-955-NeXT (outside the U.S. call +1-415-424-8500) to speak to a NEXTSTEP Technical Support Technician. If your site has a NeXT support contract, your site's support contact must make this call to the hotline. Otherwise, hotline support is on a pay-per-call basis. Thanks for using NeXTanswers! _________________________________________________________________ Written by: Eric P. Scott ( eps@toaster.SFSU.EDU ) and Scott Anguish ( sanguish@digifix.com ) Additions from: Greg Anderson ( Greg_Anderson@afs.com ) Michael Pizolato ( alf@epix.net ) Dan Grillo ( dan_grillo@next.com )
From: Mon-Sen Yang <myast2+@pitt.edu> Newsgroups: comp.sys.next.programmer Subject: Where to start? Programming for NeXT. Date: Sun, 28 Sep 1997 03:13:26 -0400 Organization: University of Pittsburgh Message-ID: <Pine.GSO.3.96L.970928031059.24288B-100000@unixs3.cis.pitt.edu> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII What books are good for learning how to program under NeXT? Know C but haven't tried Objective C yet. Your commands please. batmon
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <60m67s$4th@mtinsc03.worldnet.att.net> Control: cancel <60m67s$4th@mtinsc03.worldnet.att.net> Date: 28 Sep 1997 18:04:46 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.60m67s$4th@mtinsc03.worldnet.att.net> Sender: DHEA SEX LIKE 18<hart@sunbell.com> Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: Virtual Storefronts<webmaster@virtualstorefronts.com> Newsgroups: comp.sys.next.programmer Subject: Web Hosting Date: 28 Sep 1997 18:49:58 GMT Organization: AT&T WorldNet Services Message-ID: <60m8sm$4te@bgtnsc02.worldnet.att.net> Check it out at http://virtualstorefronts.com Reseller accounts cost $12.50/month!!! WWW.YOURNAME.COM $18.95 150MEG Disk Space Domain Name Registration Unlimited Hits - NO TRAFFIC FEES * Unlimited Email Aliases & Forwarding Unlimited Email Autoresponders Secure Server Access included 3 POP Email Accounts CGI-bin - with tons of FREE scripts! shopping carts, guestbooks, order forms Anonymous FTP/Telnet Account Redundant High Speed Connection to Internet Backbone. Five T1 lines! Free Traffic Analysis Software Installed Java, PERL, C++, Python, supported mSQL database supported! Frontpage supported - no extra charge! Internic Registration Included* *Internic will bill you $100 Activated within 24 hours! (Mon-Fri) NO SETUP CHARGES!
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <60m8sm$4te@bgtnsc02.worldnet.att.net> Control: cancel <60m8sm$4te@bgtnsc02.worldnet.att.net> Date: 28 Sep 1997 18:50:09 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.60m8sm$4te@bgtnsc02.worldnet.att.net> Sender: Virtual Storefronts<webmaster@virtualstorefronts.com> Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: ericet@CAM.ORG (Eric Tremblay) Newsgroups: comp.sys.next.programmer Subject: How many characters in myScrollView??? Date: 28 Sep 1997 19:13:32 GMT Organization: Communications Accessibles Montreal, Quebec Canada Message-ID: <60ma8s$4lt@tandem.CAM.ORG> An OpenStep question. I'm trying to figure out how many characters I have in myScrollView. I just can't seem to get it to work. Any help would be very welcomed. Here's the converted OpenStep code: (Which does NOT work) - textLengthTest:sender /* The number of characters in the Text object . */ { int HowLongIsTheText; document = [MyScrollView documentView]; /* assigns the length of the text in MyScrollView */ HowLongIsTheText = [[document text] length]; /* Displays how many characters is in the text */ [theTextLength setIntValue:HowLongIsTheText]; } Here's the original NEXTSTEP code: - textLengthTest:sender /* The number of characters in the Text object . */ { int HowLongIsTheText; document = [MyScrollView docView]; /* assigns the length of the text in MyScrollView */ HowLongIsTheText = [document textLength]; /* Displays how many characters is in the text */ [theTextLength setIntValue:HowLongIsTheText]; return self; } -- Eric "E.T." Tremblay E-Mail: ericet@cam.org (NeXTMail - MIME - ASCII) OpenStep and Rhapsody Software Design, Apple Enterprise Alliance Partner Homepage: http://www.cam.org/~ericet
From: Mon-Sen Yang <myast2+@pitt.edu> Newsgroups: comp.sys.next.programmer Subject: Question: WOF and MS Access. Date: Sun, 28 Sep 1997 20:13:24 -0400 Organization: University of Pittsburgh Message-ID: <Pine.GSO.3.96L.970928201014.24589G-100000@unixs2.cis.pitt.edu> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi, Try to use WOF NT version with MS Access. Able to access the data from Access and make HTML file, but not able to let user input data from web page and store into Access database. Is there any on-line, Webpage, or books talk about WOF and Access? batmon
From: mow@navigator.de (Markus Wenzel) Newsgroups: comp.sys.next.programmer Subject: tcgetattr in 4.2? Date: 28 Sep 1997 11:14:56 GMT Organization: Navigator Message-ID: <60le7g$23j$1@radjah.navigator.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi, I tried to compile crack with Openstep 4.2. The linker stumbled over the TERMIOS functions tcgetattr and tcsetattr while building libdes.a. However, I already have a working libdes.a built with Nextstep 3.3. So tcgetattr must work somehow with the compatibility libs, but I did not find the trick to get libdes compiled once again. Where's the lib I have to include manually? -- Navigator Markus Wenzel info@navigator.de IT Consulting & System Administration http://www.navigator.de/
From: windy@physics.utexas.edu (Windy Hardaway) Newsgroups: comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.sys.mac.programmer.misc,comp.sys.next.programmer,comp.sys.mac.system,comp.sys.mac.misc Subject: Re: No more PowerPCs (Was Re: GET RID OF THE ------- BOARD!! Date: Sun, 28 Sep 1997 20:49:27 -0500 Organization: The University of Texas at Austin, Austin, Texas Message-ID: <windy-ya02408000R2809972049270001@newshost.cc.utexas.edu> References: <see-below-0509970026400001@ip64.mountain-view2.ca.pub-ip.psi.net> <B03540AE-CE02A@206.165.44.43> <01bcba15$447c2f00$48e04382@broennic.online.no> <LTaylor7-0509971102440001@vic-ca1-02.ix.netcom.com> <don_arb-0509971214210001@sea-ts1-p59.wolfenet.com> <5vsss4$3u2$1@news.xmission.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit <The company that I used to work for employed around a thousand <technicians and customer support people, and we all depended on our <computers heavily. One of our biggest complaint for years was the <network being too slow; all our apps and shared data was stored on the <network server. Finally, when management decided to budget the money <for it, we had all our computers upgraded with larger hard drives and <the applications were installed locally. It made quite a big <difference. Shared data and important files were still kept on the <server, but the network wasn't bogged down as much when people opened <their daily programs. < <Just my 2¢. A good network infrastructure is needed for NC's, but then a having a fast reliable network is a good long term investment anyway, one that is more scalable and less prone to obsolence than investment in personal computers. Network problems are also much easier to troubleshoot (unless you use thinnet) than software problems. Add fast reliable servers and each seat would have minimal support costs. Basically NC's are glorified X-servers, where CPU power is also "served" from them. X-servers have very little support costs, and so it can be with NC's. Replace the X interface and X apps with something more familiar and friendly, e.g. Mac OS or Windows apps, and you have a winner. Tai
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <342f99ca.0@news.inreach.com> Control: cancel <342f99ca.0@news.inreach.com> Date: 29 Sep 1997 12:52:20 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.342f99ca.0@news.inreach.com> Sender: as09r04e@jflkasdfjsd.com Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
Message-ID: <3428485A.55F4@ibm.net> Date: Tue, 23 Sep 1997 23:53:15 +0100 From: Ralph Paul <repaul@ibm.net> MIME-Version: 1.0 Newsgroups: comp.sys.next.programmer Subject: Re: Is Java Better than Objective-C? References: <3414041A.E71BFBCD@online.disney.com> <B042DB4F-78B2@141.214.128.36> <1997Sep15.212941.1550@online.de> <5vndmv$2bl$1@owl.slip.net> <1997Sep17.201838.942@online.de> <5vrall$ghn1@castor.cca.rockwell.com> <604k9g$m47$1@owl.slip.net> <60647j$mj1@castor.cca.rockwell.com> <606qvo$855$1@owl.slip.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Steve Dekorte wrote: > > Erik M. Buck <embuck@palmer.cca.rockwell.com> wrote: > > In <604k9g$m47$1@owl.slip.net> Steve Dekorte wrote: > > > Uh, this may be a stupid question, but is it possible to write and > > > Objective-C > > > interpreter in ADA? If so how provable would what it interpret's be? Extend > > > these thoughts to "provability" problem in general. > > > thoughts to "provability" problem in general. > > > > > It would not matter if it was implemented in machine language. It is the > > algorithm that is difficult to prove. The language is irrelevant > > Exactly. And if you need to do something dynamic, you either use > well tested libraries (like the Objective-C runtime) or you write your > own. This still does not solve the problem you face when you have to write software that needs to be certified according to the standards applied to avionics or flight control software Anyway. BTW: What would be the point in writing an Objective-C interpreter in Ada ? What kind of proof would that be ? That I know Ada and Obj-C well enough to write a lexer/parser and some sort of virutal machine ? The proofs that need to be done are more like : prove/show that a function y = f(x,u) with the inputs u and states x produces the specified outputs y and therefore performs as intended by the formal specification ( block diagram, techincal drawings, technical documentation). Along the way you to have show program flow, control flow, data flow, .... The Obj-C runtime is not comparable. Or has anybody shown that the reliability is better then 10^-9 (:-). I think we are drifting to far away from the Obj-C <-> Java stuff. CU/2 Ralph Paul
From: Thomas Harte <tph1001@CL.cam.ac.uk> Newsgroups: comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin Subject: PGCC: has anyone compiled it? Date: 29 Sep 1997 15:43:00 GMT Organization: Cambridge University Message-ID: <60oia4$lh1$1@lyra.csx.cam.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii I have heard that there are difficulties in compatibility between gcc >= 2.7 and NS 4.1. PGCC: has anyone compiled it for NEXTSTEP 4.1? I'm running a PPro. I don't have time at the moment to install gcc (the version which came with the distribution is 2.5.8!) let alone pgcc. Just wondering if anyone has pgcc or gcc >= 2.7 binaries knocking about that are compatible with NEXTSTEP 4.1? -- thomas harte @ computer laboratory, cambridge university; tph1001@CL.cam.ac.uk; phone: MIME & NeXT Mail OK.
From: arti@lava.DOTnet (Art Isbell - remove "DOT") Newsgroups: comp.sys.next.programmer Subject: Re: Crash trap testers needed Date: 29 Sep 1997 21:15:34 GMT Organization: LavaNet, Inc. Distribution: world Message-ID: <60p5pm$pde@mochi.lava.net> References: <60fdu9$fbv@mochi.lava.net> arti@lava.DOTnet (Art Isbell - remove "DOT") wrote: > I need some assistance in testing the latest incarnation of Julie > Zelenski's ObjectError (described below). I plan to post it at archive sites > once I'm fairly certain that it works well. > > I've posted source for a beta version at ftp.lava.net/users/arti > that's available via anonymous FTP. I would appreciate those of you > interested in using a crash trap with your projects testing OEObject and > providing feedback. Holger Hoffstaette asked whether an example program was included as was the case with OEObject's predecessor, HKCrashTrap. I thought that this was a great idea, so I shamelessly ported the HKCrashTrap demo app, CrashTrap, and have included it with the OEObject distribution. This porting exercise uncovered a couple of problems with OEObject 1.2. I lifted OEObject from an OSE/NT framework. Thanks to the brilliance of the design of Microsoft's linker, when an extern symbol is referenced, its declaration depends on whether the file is in a main project or is loaded from a framework. So if you use OEObject 1.2 in a main project under OSE without changing the declaration of extern symbols, it will probably fail to work. If you're lucky enough to still be working under Mach, you probably won't experience this problem. Get OEObject 1.3 to explain this problem more thoroughly. OEObject attempts to skip its own stack frames in its backtrace since they aren't really very interesting. I discovered that different type of crashes involved different numbers of OEObject stack frames, so OEObject 1.2 actually skips 1 non-OEObject stack frame for a certain type of crash. This has been fixed in OEObject 1.3. After looking back at HKCrashTrap again, I decided to offer 2 choices for crash recovery. OEObject 1.2 requires that NSApplication's run method be overridden in order to continue after a crash. This is handy because it can also provide a means of preventing a crash from occurring when one of OPENSTEP's many new exceptions is raised but not caught. But for those who prefer the "self-contained" approach used by HKCrashTrap in which no run override is required, OEObject 1.3 provides this as an option. So pick up OEObject 1.3 at ftp.lava.net/users/arti. Its CrashTrap demo app will allow you to experience what it can do without modifying any of your projects. -- Art Isbell NeXT/MIME Mail: arti@lavaDOTnet Trego Systems (for whom I don't speak) Voice/Fax: +1 808 394 0511 OPENSTEP/NT Voice Mail: +1 808 394 0495 managed care solutions US Mail: Honolulu, HI 96825-2638
From: Cosmo Roadkill <cosmo.roadkill%bofh.int@rauug.mil.wi.us> Newsgroups: comp.sys.next.programmer Subject: cmsg cancel <3430270b.0@news.inreach.com> Control: cancel <3430270b.0@news.inreach.com> Date: 29 Sep 1997 22:47:21 GMT Organization: BOFH Space Command, Usenet Division Message-ID: <cancel.3430270b.0@news.inreach.com> Sender: 8ujafdz@asosiuwore.com Article cancelled as EMP/ECP, exceeding a BI of 20. The "Current Usenet spam thresholds and guidelines" FAQ is available at http://www.uiuc.edu/ph/www/tskirvin/faqs/spam.html Please include the X-CosmoTraq header of this message in any correspondence specific to this spam. Sick-O-Spam, Spam-B-Gon!
From: "Gregory H. Anderson" <greg@afs.com> Newsgroups: comp.sys.next.programmer Subject: OS/NT and loadable bundles - memory access rules Date: Mon, 29 Sep 1997 19:32:51 -0400 Organization: Anderson Financial Systems Inc. Message-ID: <34303AA3.543F@afs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I knew I was going to regret this when I started it this morning... Does anyone know -- or have a reference to -- the offical rules about passing memory between main code and runtime-loadable modules (DLLs) in OS/NT? In particular, I am curious about access to extern-ed variables. Can DLL code access externs from the main code? Vica versa? It seems like I can't even reference 'NSApp' without causing a SIGSEGV. Also, is it OK to create an object in a DLL-contained class and pass it back to the main code? Greg
From: Chris Roehrig <croehrig@House.ORG> Newsgroups: comp.sys.next.programmer,comp.unix.internals Subject: Re: How do I get a (struct proc *) for the current user process (in NEXTSTEP 3.3)? Date: 30 Sep 1997 00:01:01 GMT Organization: The University of British Columbia Message-ID: <60pfft$j4b$1@nntp.ucs.ubc.ca> References: <60hevp$7va$1@nnrp.cs.ubc.ca> In article <60hevp$7va$1@nnrp.cs.ubc.ca> Chris Roehrig <croehrig@House.ORG> writes: > > I'm writing a device driver for NEXTSTEP 3.3 (Mach 2.5-based) > and I need to get the struct proc * for the current user process > from within an ioctl system call in order to send that process > a signal later on (via psignal). [Just making an entry for my reference database (as indexed by DejaNews :-)] This seems to work (NEXTSTEP 3.3, from inside the ioctl system call): (struct proc *) pr = active_u[0].utask->uu_procp; /* see sys/user.h */ Thanks to Bjoern Groenvall <bg@sics.se> who put me onto the u data structure. -- Chris Roehrig croehrig@House.ORG Neuroscience and Computer Science at University of British Columbia, Vancouver http://www.House.ORG/chris http://www.sns.cs.ubc.ca/chris
From: MaRK_BeSSeY@Apple.Com (Mark Bessey) Newsgroups: comp.sys.next.programmer Subject: Re: Forcing copies of objects over DO Date: 30 Sep 1997 00:21:00 GMT Organization: Apple Computer, Inc. Message-ID: <60pglc$11gg$1@news.apple.com> References: <60ebvv$m6g$1@news.apple.com> Folowing up to my own follow-up... Mark Bessey writes > Gregory H. Anderson writes > > My question: Is there a magic incantation to force the NSImage to be > > copied across the wire, with no further dependency on the originator? > > Or is this an implementation detail built into the NSImage port coding > > that I can't override easily? > > Have you tried using the "bycopy" modifier in your method declaration? > > I've never tried this with NSImage, but it certainly works for other > classes. The class to be sent bycopy needs to have initWithCoder and > encodeWithCoder methods, but most of the AppKit/Foundation classes do. I > think that you HAVE to use setProtocolForProxy: too. Looks like this is a limitation of NSImage. As has been noted elsewhere, you can use NSData to send the data from the image's representation, so this isn't a total disaster. From a quick check, it seems like several (most?) of the AppKit classes don't support being sent bycopy. On the other hand, NSArray and NSDictionary appear to encode bycopy by default, but using byref forces them to be sent byref. The documentation could certainly be more clear about which classes honor the byref/bycopy keywords when encoding. -- Mark Bessey Apple Computer, Inc. -->I DON'T SPEAK FOR APPLE<--
From: d89cb@efd.lth.se (Christian Brunschen) Newsgroups: comp.sys.next.programmer Subject: Re: Forcing copies of objects over DO Date: 30 Sep 1997 07:51:12 GMT Organization: Lund Institute of Technology, Sweden Message-ID: <60qb1g$gqb$1@news.lth.se> References: <60ebvv$m6g$1@news.apple.com> <60pglc$11gg$1@news.apple.com> NNTP-Posting-User: d89cb In article <60pglc$11gg$1@news.apple.com>, Mark Bessey <mark_bessey@apple.com> wrote: [on the subject of sending NSImage:s 'bycopy' over DO] > >Looks like this is a limitation of NSImage. As has been noted elsewhere, >you can use NSData to send the data from the image's representation, so >this isn't a total disaster. From a quick check, it seems like several >(most?) of the AppKit classes don't support being sent bycopy. > >On the other hand, NSArray and NSDictionary appear to encode bycopy by >default, but using byref forces them to be sent byref. The documentation >could certainly be more clear about which classes honor the byref/bycopy >keywords when encoding. It would, indeed, be good if the actual behaviour of the FoundationKit and AppKit classes in these regards were accurately documented; but even better if it were both that, and more flexible - ie, that this limitation on NSImage (and whatever other classes might 'suffer' from this :) be removed (ie, the ability for them to encode themselves for bycopy passing were added). I'm sure Apple will 'think different' on this :) >-- >Mark Bessey >Apple Computer, Inc. >-->I DON'T SPEAK FOR APPLE<-- // Christian Brunschen (I speak only for myself, and not even that all the time :)
From: bresink@informatik.uni-koblenz.de (Marcel Bresink) Newsgroups: comp.sys.next.programmer Subject: Re: OS/NT and loadable bundles - memory access rules Date: 30 Sep 1997 08:19:02 GMT Organization: University Koblenz / Germany Message-ID: <60qclm$es5$1@newshost.uni-koblenz.de> References: <34303AA3.543F@afs.com> "Gregory H. Anderson" <greg@afs.com> wrote: > Does anyone know -- or have a reference to -- the offical rules about > passing memory between main code and runtime-loadable modules (DLLs) in > OS/NT? In particular, I am curious about access to extern-ed variables. > Can DLL code access externs from the main code? Vica versa? It seems > like I can't even reference 'NSApp' without causing a SIGSEGV. I had the same questions and experimented a lot with loading DLLs under OpenStep. My first impression is: The dynamic linking of DLLs allows mutual access to all dynamic data structures under control of the Objective-C runtime system. A DLL can access all objects and methods from the main code and vice versa. However, the two components _cannot_ access statically defined extern variables or C functions from the other component. > Also, is it OK to create an object in a DLL-contained class and pass it > back to the main code? This seems to be no problem. Marcel --- Marcel Bresink, University of Koblenz, Institute for Computer Science Rheinau 1, D-56075 Koblenz, Germany, Fon: +49-261-9119-421 Fax: ...-497 MIME/NeXT Mail accepted --- WWW: http://www.uni-koblenz.de/~bresink
From: schack@skyler.arc.ab.ca (Brian Schack) Newsgroups: comp.sys.next.programmer Subject: Re: OS/NT and loadable bundles - memory access rules Date: 30 Sep 1997 09:26:11 -0600 Organization: Alberta Research Council, Calgary Alberta, Canada Sender: schack@skyler.arc.ab.ca Message-ID: <vbn2kuvn58.fsf@skyler.arc.ab.ca> References: <34303AA3.543F@afs.com> In-reply-to: "Gregory H. Anderson"'s message of Mon, 29 Sep 1997 19:32:51 -0400 >>>>> "Gregory" == "Gregory H Anderson" <greg@afs.com> writes: Gregory> I knew I was going to regret this when I started it this Gregory> morning... Does anyone know -- or have a reference to -- Gregory> the offical rules about passing memory between main code Gregory> and runtime-loadable modules (DLLs) in OS/NT? In Gregory> particular, I am curious about access to extern-ed Gregory> variables. Can DLL code access externs from the main Gregory> code? Vica versa? It seems like I can't even reference Gregory> 'NSApp' without causing a SIGSEGV. Yes and yes. However, you may have to go through some effort to make it work. According to the docs, sharing classes and categories is not a problem. Functions and variables are - under NT you have to declare them specially, using some typically ugly Windows incantations. If you look at NeXTanswers 2473, there's an explanation of how to share functions and variables between a DLL (framework) and outside code. It's not that well written, but it should give you an idea. You might also want to search on the string "__declspec" in NeXTanswers to see other, possibly clearer, references to the problem. Gregory> Also, is it OK to create an object in a DLL-contained Gregory> class and pass it back to the main code? I think so, but don't quote me on that. -- ---------------------------------------------------------------------------- Brian Schack |mailto:schack@arc.ab.ca | "I don't want to achieve Alberta Research Council |http://www.arc.ab.ca | immortality through my 6815 8th St NE | | work ... I want to achieve Calgary, Alberta |ph: (403) 297-7564 | it through not dying." Canada T2E 7H7 |fax: (403) 297-2339 | - Woody Allen ----------------------------------------------------------------------------
From: ezhxuw@stm544 (Werner Burri) Newsgroups: comp.sys.next.programmer Subject: Re: OS/NT and loadable bundles - memory access rules Date: 30 Sep 1997 15:13:58 GMT Organization: UBS Network Services Message-ID: <60r4vm$eni@svdns1.ubinet.ubs.com> References: <34303AA3.543F@afs.com> <60qclm$es5$1@newshost.uni-koblenz.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: bresink@informatik.uni-koblenz.de > I had the same questions and experimented a lot with loading DLLs under > OpenStep. My first impression is: The dynamic linking of DLLs allows mutual > access to all dynamic data structures under control of the Objective-C > runtime system. A DLL can access all objects and methods from the main code > and vice versa. However, the two components _cannot_ access statically > defined extern variables or C functions from the other component. > > > Also, is it OK to create an object in a DLL-contained class and pass it > > back to the main code? > > This seems to be no problem. > > Marcel > This is correct. If you program a Bundle, you can only reach the Objective-C methods of the main code. That's the reason why Next added the class method +sharedApplication to NSApplication, because you've got no chance to reach NSApp from a bundle (on NT). If you want to use C-functions in the main code and in the bundle, you have to place this code into a Framework. On OpenStep for Mach it is enough to link this Framework against the main-code. On OpenStep for NT you have to link this Framework against the main code and every bundle, which uses this functions. E.g. To bring up an NSRunAlertPanel from a bundle, you have to link AppKitt.framework against the bundle too (only on NT). The easiest way to work with global variables is to capsulat them in classes and access them only via class methods. So you will never have problems to port the code to an other platform or to split it up into multiple bundles. Christian
From: thomas@catlan.met.FU-Berlin.DE (Thomas Hensel) Newsgroups: comp.sys.next.programmer Subject: Debugging Postscript ? exceptions ? Date: 30 Sep 1997 16:18:51 GMT Organization: Freie Universitaet Berlin Message-ID: <60r8pb$d6k$1@fu-berlin.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi ! I need help with a program I have written. Sometimes it crashes randomly with this output: : App: openFile: /Users/thomas/VimDocs/5.visMap - type visMap : DPS client library error: PostScript program error, DPSContext 5667cc : %%[ Error: undefinedresult; OffendingCommand: bin obj seq, type=129, elements=3, size=28 ]%% : An uncaught exception was raised : DPS client library error: PostScript program error, DPSContext 5667cc : %%[ Error: undefinedresult; OffendingCommand: bin obj seq, type=129, elements=3, size=28 ]%% What does it mean "An uncaught exception was raised" ? How can I catch the exception with gdb or with my program ? Thanks, Thomas -- || Who: Dipl. Phys. Thomas Hensel MIKS - Meteorologische Informations- || EMail: thomas@bibo.met.FU-Berlin.DE und Kommunikations-Systeme || Voice: (+49 30) 838 71 225 an der Freien Universitaet Berlin || FAX: (+49 30) 791 90 02 Schmidt-Ott-Str. 13 - 12165 Berlin
From: tom@basil.icce.rug.dev.null.nl (Tom Hageman) Newsgroups: comp.sys.next.programmer Subject: Re: tcgetattr in 4.2? Date: 30 Sep 1997 17:41:38 GMT Organization: Warty Wolfs Sender: news@basil.icce.rug.nl (NEWS pusher) Message-ID: <EHAGvD.EB6@basil.icce.rug.nl> References: <60le7g$23j$1@radjah.navigator.de> mow@navigator.de (Markus Wenzel) wrote: > I tried to compile crack with Openstep 4.2. The linker stumbled over the > TERMIOS functions tcgetattr and tcsetattr while building libdes.a. However, I > already have a working libdes.a built with Nextstep 3.3. So tcgetattr must > work somehow with the compatibility libs, but I did not find the trick to get > libdes compiled > once again. Where's the lib I have to include manually? Gone. tcgetattr() is a POSIX function, and libposix.a was dumped in OS4.0 (due to lingering bugginess and lack of resources at NeXT at the time I guess :-/) If the des code has a BSD-tty compatibility compilation mode, you should be able to use that. Hope this helps, (Oh well...|-/) --Tom.
From: marcel@system.de Newsgroups: comp.sys.next.programmer Subject: Re: Forcing copies of objects over DO Date: 30 Sep 1997 18:34:43 GMT Organization: Technical University Berlin, Germany Distribution: world Message-ID: <60rgo3$fl6$1@brachio.zrz.TU-Berlin.DE> References: <60pglc$11gg$1@news.apple.com> In article <60pglc$11gg$1@news.apple.com> MaRK_BeSSeY@Apple.Com (Mark Bessey) writes: > Folowing up to my own follow-up... > Mark Bessey writes > > Gregory H. Anderson writes > > > My question: Is there a magic incantation to force the NSImage to be > > > copied across the wire, with no further dependency on the originator? > > > Or is this an implementation detail built into the NSImage port coding > > > that I can't override easily? > > > > Have you tried using the "bycopy" modifier in your method declaration? > > > > I've never tried this with NSImage, but it certainly works for other > > classes. The class to be sent bycopy needs to have initWithCoder and > > encodeWithCoder methods, but most of the AppKit/Foundation classes do. I > > think that you HAVE to use setProtocolForProxy: too. > > Looks like this is a limitation of NSImage. [...] Well, NSImages can contain NSCachedImageReps, which reference windows using an integer. Making sure a 'copy' of an NSCachedImageRep stays consistent looks like a fairly difficult task. Several solutions lool possible at first, but all have some not entirely obvious gotchas. I haven't looked at it in detail, but it seems like an explanation. Marcel
From: markfr@galway.cse.tek.com (Mark S Frank) Newsgroups: comp.sys.next.programmer Subject: Warning: Object compatibility? Date: 30 Sep 1997 20:28:37 GMT Organization: Tektronix, Inc. Distribution: world Message-ID: <60rndl$sbl$1@bvadm.bv.tek.com> I am trying to port an app from NS 3.3 to OS/NT. When I try to run the program from the debugger, I get warning messages like: warning: Object compatibility method 'retain' in class Controller has been executed at least once. Can anyone tell me what this means, and what I need to do to fix it? - Thanks for any help, mark.s.frank@tek.com
From: Greg_Anderson@afs.com (Gregory H. Anderson) Newsgroups: comp.sys.next.programmer Subject: Re: Warning: Object compatibility? Date: 30 Sep 1997 20:50:31 GMT Organization: Anderson Financial Systems Inc. Distribution: world Message-ID: <60romn$bfk@shelob.afs.com> References: <60rndl$sbl$1@bvadm.bv.tek.com> Mark S Frank writes > I am trying to port an app from NS 3.3 to OS/NT. > When I try to run the program from the debugger, I get > warning messages like: > > warning: Object compatibility method 'retain' in class Controller > has been executed at least once. > > Can anyone tell me what this means, and what I need to do to fix it? This is just a warning that "grandfathered" methods are being invoked. It typically happens when you load a nib file that has not been resaved in a 4.x version of IB. I see them all the time while loading old PasteUp and WriteUp documents, because those files were saved as typedstreams which contain now-obsolete classes. They do get annoying after a while, but I don't know of any way to suppress them. -- Gregory H. Anderson | "We're in the land of the blind, Visionary Ophthalmologist | selling working eyeballs, and they Anderson Financial Systems | balk at the choice of color." -- Tony greg@afs.com (NeXTmail OK) | Lovell, on Mac user reactions to NeXT
These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Marcel Waldvogel and Netfuture.ch.