*BSD News Article 61747


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!netnews.upenn.edu!dsinc!news.icubed.com!scramble.lm.com!news.psc.edu!nntp.sei.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!news.mathworks.com!news.kei.com!nntp.coast.net!howland.reston.ans.net!vixen.cso.uiuc.edu!milo.mcs.anl.gov!cmt.anl.gov!SCANDORA
From: scandora@cmt.anl.gov (Tony Scandora 708-252-7541)
Newsgroups: misc.misc,misc.forsale.computers.discussion,comp.os.os2.marketplace,comp.os.msdos.misc,comp.os.ms-windows.win95.misc,comp.os.ms-windows.nt.misc,comp.os.mach,comp.os.ms-windows,comp.os.ms-windows.advocacy,comp.os.misc,comp.os.magic-cap,comp.os.lynx,comp.os.linux.misc,comp.os.linux,comp.os.386bsd.development,comp.os.386bsd.bugs
Subject: Re: Win 95 stability
Date: 19 Feb 1996 18:46:58 GMT
Organization: Argonne National Laboratory
Lines: 58
Message-ID: <4gagj2$iit@milo.mcs.anl.gov>
References: <Pine.D-G.3.91.960124225044.20914C-100000-100000@erc.cat.syr.edu> <4eu2n6$bvu@rubens.telebyte.nl> <4fflmt$jh3@c4.hrz.uni-giessen.de> <4fhldb$2he@rubens.telebyte.nl> <4fhqo5$t6d@cegt201.bradley.edu>,<4fnbud$tq7@c4.hrz.uni-giessen.de> <4ft7pt$2s7@milo.mcs.anl.gov>,<4g1ne8$t63@c4.hrz.uni-giessen.de>
Reply-To: scandora@cmt.anl.gov
NNTP-Posting-Host: ipep.cmt.anl.gov
Xref: euryale.cc.adfa.oz.au misc.misc:30126 misc.forsale.computers.discussion:3081 comp.os.os2.marketplace:2118 comp.os.msdos.misc:48905 comp.os.ms-windows.win95.misc:105531 comp.os.ms-windows.nt.misc:92011 comp.os.mach:5067 comp.os.ms-windows.advocacy:111962 comp.os.misc:4727 comp.os.magic-cap:1511 comp.os.lynx:1716 comp.os.linux.misc:87527 comp.os.386bsd.development:3360 comp.os.386bsd.bugs:3099

In article <4g1ne8$t63@c4.hrz.uni-giessen.de>, chris.traxler@theo.physik.uni-giessen.de (Chris Traxler) writes:
>scandora@cmt.anl.gov (Tony Scandora 708-252-7541) wrote:
>
>>In article <4fnbud$tq7@c4.hrz.uni-giessen.de>, chris.traxler@theo.physik.uni-giessen.de (Chris Traxler) writes:
>>How long does it take to develop applications in VB or Access or PB or ...
>>v. GCC and curses or Tcl?
>
>In general, gcc is neither slow nor bad. It optimizes way better than
>some commercial compilers. I am not a developer of desktop apps; I am
>right now writing a numerical simulation of a field theory. 
>
For such, you can't beat gcc.  It usually generates correct code, and 
fast, too.  On the other hand, if you need a nontrivial user interface 
with intelligent popup panels and context sensitive help adaptable to the
experience and needs of many users, connected to a database server who
knows where and/or creating complex formatted output, you can develop it 
in much less time using any of many PC tools priced well < US$ 1000 
than you can with Tcl/Tk.

>You are more or less saying "one has to use DOS/Win if one wants to
>develop applications for customers who use DOS/Win". That is correct,
>but does not contribute much to our discussion about the quality of
>the various OS's.
>
I should have made that clearer.  What I meant to say was that
available applications and development tools for complex user interfaces 
and database access from companies likely to be in business next year 
overwhelmingly favor Windows.  The 16 bit Microsoft OSs are pathetic,
and the hybrid 16/32 bit Windows 95 has some serious weaknesses compared
to Linux, but application and development tool availability often
requires their use.

>I can't say a word about NT, except that I would not buy it. The
>backslash was a joke, yes, but old software? All too often, Windoze
>users think that all their problems are due to faulty "software". One
>example: Maple is stable on all platforms I've seen it EXCEPT on
>Windows. Do you think this is by chance?

The 16 bit APIs are a real disaster.  Is it WordPerfect or Windows for
Workgroups that caused the GPF?  Who cares?  The document is garbage and
you have to reboot no matter which one failed.  Unless the application
fits neatly into 64K pieces, it's a mess.  I'm never surprised to see good
software fail on a PC.  Did you ever try to constrain big arrays or lots 
of calls to malloc to the 8086 addressing scheme?

On the other hand, the Win32 APIs are not hampered by compatibility
requirements for obsolete architectures.  Compile and link a 32 bit
Maple and observe it on NT.  It might even be reliable on Windows 95.
What is still running on the majority of PCs is a disgrace, but
Microsoft got paid for it anyway.  That money was used to develop what
is finally coming out, and which works.

I hate to see any company owning the world, but Win32 is useful and
reliable, and has a large supply of good, cheap applications and
development tools.  I wish there were some competition.

Tony Scandora, Argonne National Lab, 708-252-7541
scandora@cmt.anl.gov