Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!pasteur!ucbvax!TIS.COM!dmb
From: dmb@TIS.COM (David M. Baggett)
Newsgroups: comp.sys.atari.st
Subject: Re: Standardization
Message-ID: <8803172142.AA00829@TIS.COM>
Date: 17 Mar 88 21:42:30 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 54


   Personally I think that the lack of standardization is the one thing that
has kept the ST from being major competition with the Mac.  Atari is very
much at fault for this ("ha, I told you so" really doesn't cut it), expecially
considering that they were really pushing this thing as a "more serious" or
"business" machine.  But this isn't just more complaining about the state of
atari's management, etc. etc.
   Frankly, ST software as a group is a terrible mess.  I certainly am not
criticizing the programmers for this (I am one and proud too), and I am
not saying the software runs poorly.  The problem is that, from the beginning
we have not known a damn thing about how the various OS componens work GEM,
TOS, AES, Bios, etc.  Much of the information has been discovered through
playing around -- hacking -- and has been passed through bulletin boards and
interest groups.  Remember the "fine" Developer's Package?  It was incomplete,
really unclear (and ambigous), and worse, frequently incorrect.  That started
developers off on the wrong foot in the first place.  The Abacus books helped
but they were vague and occasionaly wrong too.  Rember the article in STart
about message passing between desk accessories (they used DEGAS as an 
example)?  That's exactly the kind of thing developers should have known
about and supported all along.
   The point of all this raving is that, since everyone knew very little about
the machine and its OS initially, and since Atari never set forth any standards
to speak of, every developer found his own way of doing things.  The result:
Incredible lack of portability and compatibility between applications. 
   This may sound like whining, but think about this:  It is currently 
impossible to write a "Juggler" program that works on all ST software (by
"Juggler" I mean a program that allows you to have more than one program
in memory and to switch between them).  In fact, it's impossible to write
one that works on much software at all (Michtron's attempt only works with
straight GEM applications -- forget NeoChrome, Spetrum 512, and all those  
other programs with hardware hacks (like HBL's etc.))
   This is because there is no standard way of doing anything.  No one told
developers that they shouldn't use HBL's.  No one said not to mess with the
mouse handler.  Everyone did these things (and who can blame them), but by
doing so any hope of upward compatibilty and portability was tossed out the
window.
   Compaare this situation to the Mac.  That machine has evolved so much,
and all the software still works.  (Examples: The new multifinder which 
does "fer real" multitasking on EXISTING applications, the 68020 upgrades
which cause no problems, upward compatability with the new Mac II's)
   The ideal:  A really good toolbox (like the Mac's) which would give
programmers good routines so they wouldn't have to hack their own in 
gross, nonportable 68000 asm.  I've been working on an ST animation program 
for over a year, and I just recently looked at the mac toolbox (so I'm not
just some MacJunkie trying to sell you on my machine -- I don't even own 
one).  But wow.  If we'd been blessed with a system like QuickDraw, you bet 
this machine would have blown the mac and amiga out of the water.
   So I think standardization is essential.  It's probably too late for
the ST (at least the GEM ST -- there's still hope for a new OS), but maybe
Atari will eventually learn not to throw things on the market without good
documentation.  Hardware is useless without software and software is useless
without documentation.
	:-(
				Dave Baggett
				dmb@tis.com