This is Various-comments-questions-suggestions in view mode; [Up]
Date: Sun 12-May-1989 20:31:32 From: Unknown Subject: Various comments/questions/suggestions Here is a list of some comments, suggestions, and problems I've been having. I'm really glad to see you NeXT people (Ali and Avadis especially) reading this newsgroup and responding! Everybody else, too, this is a very informative and relatively flame free forum, let's keep it up!! 1)Click to type vs. mouse position focus (ala Sun). I want a choice! This should be settable! Also, it's very anoying when you hide something, and then NOBODY is the listener. You could use a stack paradigm; App A is listening, then you bring up App B, now App B is listening. Bring up App C, and it listens. Hide App C and App B gets the focus. Quit out of App B and App A gets the focus again. This too could be settable by the user. In the preferences panel, a new 'focus' pallette would be needed for these two settings. 2)User settable keybindings. I want to be able to make everything answer to emacs commands (as best as possible). I'll do the work if applications let the keys be rebound. Even X let's you do that! 3) CONSISTANCY!! This is VERY important! You've done a good job mostly, (command x,c,v are the same everywhere), but Define is c-d usually, in Edit c-d is something else (Find Previous), in somethings you have to use the menu because define is not keyboard shortcutted (Edit), etc. The Mac really stressed the ease of use; once you know ONE application, the rest are easy because all the keybindings and menu layouts are the same (as much as possible). PLEASE try to do this, it really helps! Any word yet on getting some more fonts? I assume Adobe is working on it (we use Optima a lot here in Kodak!) 1) lpr <file> cuts off text on the right. When you bring it into Edit and Print it comes out OK. PRINTING (from applications): Many 'beta' applications fry PS when you try to print (FrameMake, sometimes WriteNow). Laurie Sefton gave me a good hack, when the PS gets bugged log out and log into the 'exit' account (no passwd); it will reset PS and log you out. Log back into your account and things get better. There should be some code to prevent bad applications from ruining PS for everything else! If you're trying to use yellow pages logins (ie, the +:0::0.... stuff at the end of /etc/passwd) and your home path is somewhere else (like on a Sun) and there is no .NeXT directory in your home path, you can't log in. I assume it's looking for your dock info, etc. How about a message, so you know what's going on? "Sorry no dock info" or something. Also, the /etc/fstab file says it is not read if you're using netinfo. that's not true!! I have some entries that mount my old sun's disks in /etc/fstab, and whenever I reboot, it reads those and mounts my disks! (I realize it's still young, here's some improvement suggestions) 1) No options for inserting file, and/or saving a letter as a file. I do this a lot! 2) No 'grouping' commands. It's very nice to be able to 'mark' messages, then forward them all to someone, delete them all, etc. See the Andrew Message system. 3) The 'reply to' field isn't quite right. The problem has to do with domains, I believe. When I send a message out, it looks like you can reply to me with 'rose@onyx' (onyx is the CUBE) or 'rose@quartz' (my Sun) depending on where I sent it from. Nobody in netland knows about quartz or onyx, sice their local names. I don't know how mail would know, but people have told me they couldn't 'reply'. Maybe I could tell mail to put a specified string in the 'reply to' area and I can give it my uucp address (...ll-xn!atexnet!rose)? 1) When you 'Open Directory' the view should be the same as the current open window. It defaults to an icon view. INTERFACE BUILDER: Here's a real can of worms! I love the concepts, and think there is a lot of good work in it already, but the user interface, and method of operation needs some more work. It's waaaaaaaay too confusing and cluttered. By the time you open up all the windows and palletes, it looks like you're running six applications! It's also not at all intuitive how you hook things up and examine conections. I took the developers course and that helps a lot, but the docs we have are all for the .8 IB, and .9 is significantly changed. How about shipping the .9 IB docs with the system, or at least posting ascii versions here? The up to date docs should help a lot. One suggestion is with the sounds/icons palettes. Why not combine them into a 'resources' palette and choose between sounds, icons, and whetever else. Alternatively, maybe they could go in the 'Pallettes' menu (add two more choices at the top: sound and icons). Also, with these 'resources', there is a two way switch for system and local. This is nice, but you really need three switches; system, local (to the user), and program only. The local is just that program you're building. There should be a user local so something you want to put in all your applications (like your picture!) will be in one convienient place. Another problem with sounds is their size. When I import a sound there should be a way to make a link instead of a copy. I know there is a chance the original will go away, but the programmer can make the choice and deal with the consequences!! Another thing you need is 'Close Application' for when you're done working on something to get it and its' menus out of the way. I haven't used it enough to more suggestions, anyone else? 1)Has anyone heard of (and know where I can get) some software to 'read' a text file? i.e. convert ascii to voice. DEC has a product that does a reasonable job, called DecTalk, you hook it up like a terminal (RS232 port) and send it text. It talks it out a speaker. The NeXT machine would be an ideal platform to do this on!! 2) When you bring up an application, sometimes it's menu in the upper left just sits on top of the menu for the previous application! It's hard to reproduce, but it definetely happens sometimes! 3) Minor bug in more: when it's at the page and asks for '--More-- (n%)', the --More-- string stays on the screen. It should go away. 4) G++ and/or C++. There's been discussions about this in gnu.g++, comp.lang.c++, and here in comp.sys.next. I know all the IB stuff and most else has the objective-c wrapper around it, but some of us have lots of code written in c++. I think people at NeXT should take the responsibility of either getting g++ to compile on the cube (and then including it in 1.0), or at least verifying that at&t cfront compiles and runs on it. MANY people have asked me about this here (since we do lots of c++). 5) What happened to Lisp? I though we were supposed to get a terminal based version with .9? I can't find it. >From: jgreely@previous.cis.ohio-state.edu (J Greely)
Date: Sun 13-May-1989 17:31:09 From: Unknown Subject: Re: Various comments/questions/suggestions In article <238@quartz.atexnet.UUCP> rose@quartz.UUCP (Robert Rose) writes: >Here is a list of some comments, suggestions, and problems I've >been having. Thanks. It's nice to hear someone else's candid opinions. I'm really glad to see you NeXT people (Ali and >Avadis especially) reading this newsgroup and responding! Unofficially, of course :-) >GENERAL: >1)Click to type vs. mouse position focus (ala Sun). I want a choice! This is a religious issue. In our environment, we push X and NeWS, so most people think of c-t-t as a bad thing. It isn't, necessarily, but I find the frustration factor is rather high among people who are used to X. >This should be settable! Also, it's very anoying when you >hide something, and then NOBODY is the listener. Better yet, if you get a "printer out of paper" dialog, and click to make it go away, you've lost your active window. >In the preferences panel, a new 'focus' pallette would be needed >for these two settings. Actually, if you take the "desktop" model far enough, you could state that it's intuitive that removing the top ("active") document automatically sets your focus on the new top document (which sounds suspiciously like that M-word). That would be acceptable as a default, and then allow the other option to be "focus follows mouse". I think this would make most of our people happy. Those comfortable with the Mac would understand how the click-to-type worked, and the X people could feel at home. The NeWS people would still be disgusted because it's not *their* PostScript :-) >FONTS: >Any word yet on getting some more fonts? I assume Adobe is working >on it (we use Optima a lot here in Kodak!) Adobe did announce that they were working on getting their entire font library ready for use on NeXT machines, presumably in the same price range they use for the PC and Mac worlds. >LPR: >1) lpr <file> cuts off text on the right. When you bring it into >Edit and Print it comes out OK. What happens if you use enscript? And what do you mean by ok? Is it that files too wide to be printed get chopped off from lpr, and wrapped from Edit and Print? I consider this acceptable, since lpr is only doing what I told it to. Of course, to conform in our environment, I'll be replacing /usr/lib/transcript/pstext with a wrapper for enscript (more reliable, and easier to reconfigure). This will also make long lines wrap, but I can live with it :-). >PRINTING (from applications): >Many 'beta' applications fry PS when you try to print (FrameMake, sometimes >WriteNow). Well, FrameMaker doesn't appear to print through lpd, so it's *got* to be doing something wrong. (print from Frame with not enough paper; you'll never see an "out-of-paper" dialog box, and lpd will think all is well) But I've never had a problem with the window system getting scrogged, just the printer. What exactly happens to you? > Laurie Sefton gave me a good hack, when the PS gets bugged >log out and log into the 'exit' account (no passwd); it will reset >PS and log you out. Sort of. According to the latest I've heard, the way to keep npd running is to make sure that everyone has "public window server" set in Preferences. The 'exit' hack works, but this supposedly avoids the problem altogether. Still a bug, though. >If you're trying to use yellow pages logins (ie, the +:0::0.... stuff at >the end of /etc/passwd) and your home path is somewhere else (like on >a Sun) and there is no .NeXT directory in your home path, you can't log in. I can't duplicate that here. Under 0.9, I haven't had anyone be unable to log in due to lack of a .NeXT directory. Can you give me a precise repeat-by? >Also, the /etc/fstab file says it is not read if you're using netinfo. >that's not true!! I have some entries that mount my old sun's disks >in /etc/fstab, and whenever I reboot, it reads those and mounts my disks! Actually, it's not read unless you enable YP (see the manual for NetInfo). I'm jealous, because I couldn't get it to do that, and ended up shoving my NFS mounts in NetInfo (which, with documentation, is pretty easy). Part of my problem is that /etc/mount dumps core whenever I try to use it with the -a option. >MAIL: >1) No options for inserting file, and/or saving a letter as a file. >I do this a lot! I don't use Mail, but this would be a real problem if I did. If the NeXT Mail program is less capable than standard Berkeley mail, I won't touch it. Friendly user interfaces are one thing (and BSD mail needs one), but sacrificing features for tail-wagging is another. All heat aside, it *is* a young program, and needs to be solid for the basics first, but I do hope someone is working on making it more capable. >3) The 'reply to' field isn't quite right. The problem has to do with >domains, I believe. When I send a message out, it looks like you can >reply to me with 'rose@onyx' (onyx is the CUBE) or 'rose@quartz' >(my Sun) depending on where I sent it from. How do you use the NeXT mailer from your Sun? If both machines generate bogus reply-to's, the problem isn't in Mail. I'm not sure what's wrong here. This is one of those cases where the headers would be useful (so, if you sent me sample mail from both, I might be able to see what's going on. Doesn't mean I can tell you how to fix it, but at least I should (he said, *should*) be able to tell you where to start looking). >INTERFACE BUILDER: > How about shipping the .9 IB docs with >the system, or at least posting ascii versions here? The up to date >docs should help a lot. Funny, I thought they *were* shipped, online. The large document in the Librarian that contains the whole preliminary chapter (Chapter 7, BTW). If you can't get a successful search for it in the Librarian, try firing up WriteNow and opening /NextLibrary/Documentation/NeXT/SysRefMan/07_IntfBuilder.wn Runs to about 70K of text, with twelve illustrations. I don't have mine here, or I'd even give you a page count. >OTHER: >1)Has anyone heard of (and know where I can get) some software to 'read' a >text file? i.e. convert ascii to voice. Not listed in the third-party notes I have, but I'd be surprised if someone doesn't do it fairly soon. >2) When you bring up an application, sometimes it's menu in the upper left just >sits on top of the menu for the previous application! It's hard to >reproduce, but it definetely happens sometimes! If the former active App is busy doing something (like Mathematica running a long calculation), this will happen. I'd blame it on applications that aren't completely conformant with the interface guidelines yet. >3) Minor bug in more: when it's at the page and asks for '--More-- (n%)', the --More-- string stays on the screen. It should go away. This is a hacked more. Someone has "done things" with it. I suspect the modification came from CMU. It doesn't bother me, since I use less (shipped as standard software, no "less"). >I know all the IB stuff >and most else has the objective-c wrapper around it, but some >of us have lots of code written in c++. I think people at NeXT >should take the responsibility of either getting g++ to compile >on the cube (and then including it in 1.0), or at least verifying >that at&t cfront compiles and runs on it. Some of us have lots of code written in COBOL, too. Should *that* be standard software? Not trying to jump on you, but why should NeXT support two completely different object-oriented extensions to C, one of which they never use? You lost me here. > MANY people have asked me >about this here (since we do lots of c++). So pick up G++ and see if it compiles. >5) What happened to Lisp? I though we were supposed to get a terminal >based version with .9? I can't find it. It works fine. I haven't gotten any real feedback from our local lisp hackers yet, but it's certainly there. The manual pages are incorrect, but the release notes are right. Script started on Sat May 13 13:24:01 1989 % cl Allegro CL 3.1.0.5 [NeXT] (4/2/89 3:43) Copyright (C) 1985-1989, Franz Inc., Berkeley, CA, USA <cl> (exit) ; Exiting Lisp % exit script done on Sat May 13 13:24:13 1989 -=- J Greely (jgreely@cis.ohio-state.edu; osu-cis!jgreely) >From: wsd@cs.brown.edu
Date: Sun 13-May-1989 18:13:49 From: Unknown Subject: Re: Various comments/questions/suggestions In article <238@quartz.atexnet.UUCP> rose@atexnet.UUCP (Robert Rose) writes: GENERAL: 1)Click to type vs. mouse position focus (ala Sun). I want a choice! Part of the NeXT interface, as on the Mac, is the notion that moving the mouse does nothing. Clicking performs actions. It is very important to be completely consistent. 2)User settable keybindings. I want to be able to make everything answer to emacs commands (as best as possible). I'll do the work if applications let the keys be rebound. Even X let's you do that! What do you mean? As far as I know (which isn't much), X let's you globally redefine keys, but how does one set which keys do what in different applications? If you want everything to look like emacs, then use emacs for everything (I do!). 3) CONSISTANCY!! This is VERY important! You've done a good job mostly, (command x,c,v are the same everywhere), but Define is c-d usually, in Edit c-d is something else (Find Previous), in somethings you have Yes, consistancy is critical, but application specific keystrokes cannot be standardized. There is a fuzzy boundary between commands that all NeXT apps will support and should therefore be standard, and commands which most apps will support and should maybe be standard. NeXT made these decisions, perhaps some more standards will evolve (as they did on the mac). FONTS: Any word yet on getting some more fonts? ... yes, more fonts is always better. Many 'beta' applications fry PS when you try to print (FrameMake, sometimes ... what do you mean "PS"? My personal wish: A DPS implementation that performs anti-aliasing, especially of fonts. With that beautiful gray-scale screen and outline fonts, you have the perfect opportunity. It is sooo easy to generate the bitmaps for the fonts at high resolution and filter them down. Then drawing (just a blit) to the screen is just as fast, things just look twice as good. - - - - - - - - - - Scott Draves | Space... The Final Frontier wsd@cs.brown.edu | uunet!brunix!wsd wsd@browncs.bitnet Box 2555 Brown U Prov RI 02912 >From: eht@cs.cmu.edu (Eric Thayer)
Date: Sun 15-May-1989 05:48:15 From: Unknown Subject: Re: Various comments/questions/suggestions >>PRINTING (from applications): >>Many 'beta' applications fry PS when you try to print (FrameMake, sometimes >>WriteNow). > >Well, FrameMaker doesn't appear to print through lpd, so it's *got* to >be doing something wrong. (print from Frame with not enough paper; >you'll never see an "out-of-paper" dialog box, and lpd will think all >is well) > > But I've never had a problem with the window system getting >scrogged, just the printer. What exactly happens to you? FrameMaker prints by sub-classing the view class versus spooling a PostScript file and forking off an lpr. Which means npd is used instead of lpd. This is the recommended way of printing. The only fatal bug in printing reported occurrs if you leave object handles displayed while printing. I have a fix for this but I'm going to wait and post it when I have patterns printing. -greg. >From: steve@acorn.co.uk (Steve "Daffy" Hunt)
Date: Sun 14-May-1989 21:41:06 From: Unknown Subject: Re: Various comments/questions/suggestions In article <WSD.89May13141349@transit.cs.brown.edu> wsd@cs.brown.edu writes: >In article <238@quartz.atexnet.UUCP> rose@atexnet.UUCP (Robert Rose) writes: [...] > > 2)User settable keybindings. I want to be able to make everything > answer to emacs commands (as best as possible). I'll do the work if > applications let the keys be rebound. Even X let's you do that! > >What do you mean? As far as I know (which isn't much), X let's you >globally redefine keys, but how does one set which keys do what in >different applications? If you want everything to look like emacs, >then use emacs for everything (I do!). Nooooo! Way off beam... The X Toolkit's translation mechanism lets the user specify key bindings (or more accurately event bindings, because you can rebind mouse buttons and motion etc too) on a specific basis (naming one particular widget in a particular application) or a general basis (say, all the Text widgets in a particular application) or globally (all the Text widgets in all applications). Or anything in between, like all the Text widgets in some subwindow of a particular application. Mind you, how useful this all is depends on how well thought out the available translations are, and how well thought out the hierarchical naming scheme for all those widgets is. That depends on the widget writer and the application writer. Documenting it is a bit of a nightmare too. Steve. use their mouse for anything more important than selecting which vt100 emulator to type at? :-) >From: eht@f.word.cs.cmu.edu (Eric Thayer)
Date: Sun 17-May-1989 07:17:17 From: Unknown Subject: Re: Various comments/questions/suggestions In article <238@quartz.atexnet.UUCP> rose@quartz.UUCP (Robert Rose) writes: > >Here is a list of some comments, suggestions, and problems I've >been having. I'm really glad to see you NeXT people (Ali and >Avadis especially) reading this newsgroup and responding! >Everybody else, too, this is a very informative and relatively >flame free forum, let's keep it up!! > > >MAIL: > >(I realize it's still young, here's some improvement suggestions) > >1) No options for inserting file, and/or saving a letter as a file. >I do this a lot! In 0.9 you can include a file in a mail message simply by dragging it from the browser into the mail message window. You can even drag directories if you want! (As Avie suggested, try dragging "/" in - I can see the Internet clogging already!) This feature is similar to that nifty bit of Paul's draw program that lets you drag PS and TIFF files into the drawing window. Neat! >INTERFACE BUILDER: > >Here's a real can of worms! I love the concepts, and think there is a lot >of good work in it already, but the user interface, and method of operation >needs some more work. It's waaaaaaaay too confusing and cluttered. By the >time you open up all the windows and palletes, it looks like you're >running six applications! It's also not at all intuitive how you >hook things up and examine conections. I totally concur about IB - it is so amazing what it can do but keeping track of everything can be a nightmare. Will there ever be some sort of graphical net-like structure to view the connections between objects? Why is the objects window so small? And I must admit that I really don't like the two way switch. In order to flip the switch you click on the gray option. To me, something gray has always meant something disabled (sorry, Macintosh bias showing...) Where are the interface guideline thought police? :) >One suggestion is with the sounds/icons palettes. >Also, with these 'resources', there is a two way switch for system >and local. This is nice, but you really need three switches; system, >local (to the user), and program only. The local is just that >program you're building. There should be a user local so something >you want to put in all your applications (like your picture!) will >be in one convienient place. Another good suggestion! I am still struggling with trying to figure out a way to move sounds and icons I have installed in one nib file to another. I can't get it to copy/paste between nib files :( julie zelenski stanford university >From: eht@cs.cmu.edu (Eric Thayer)
These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Marcel Waldvogel and Netfuture.ch.