Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.csd.uwm.edu!uakari.primate.wisc.edu!nic.MR.NET!umn-cs!hall!rosenkra
From: rosenkra@hall.cray.com (Bill Rosenkranz)
Newsgroups: comp.sys.atari.st
Subject: Re: ROM disassembly for TOS 1.4
Message-ID: <4630@hall.cray.com>
Date: 9 Sep 89 02:55:33 GMT
References: <9401@chinet.chi.il.us> <1666@atari.UUCP> <71359@yale-celray.yale.UUCP> <1677@atari.UUCP> <4576@hall.cray.com> <21985@cup.portal.com>
Reply-To: rosenkra@hall.UUCP (Bill Rosenkranz)
Organization: Cray Research, Inc., Mendota Heights, MN
Lines: 77

In article <21985@cup.portal.com> Bob_BobR_Retelle@cup.portal.com writes:
=i said:
=>just think how much more "friendly" atari would seem to developers (over,
=>say, commodore) if they say "hey, here is the full source on disk, guys".
= 
=Drop by your local B.Dalton Booksellers or Waldenbooks book stores, and take
=a look at the *multi-volume* set of Amiga documentation that's available to
=ANYONE (without having to pay a $300 "initiation fee")

do these include src code? i thought not.

actually, when i thought about switching development platforms, i priced
these and it amounted to something like $120 or so. add $200 for a decent
compiler (i don't know amiga s/w prices so don't flame...this is just a
guess) and you come upto atari's $300 mark. since i got the dev kit way back
when for $200 (from xanth), it seemed like a pretty good deal (compiler,
docs, reduced prices on h/w, consultation, etc). i STILL use alcyon,
despite it's many shortcommings (except, of course apratt's aln which is
an excellent linker). 

=Then ask yourself what is Atari Corp trying to do by keeping all this info
="secret"...?

good question. the answer, of course, lies in profit margins. atari didn't
(and still doesn't) need to do much to launch the ST. the "community",
especially in germany, keeps it afloat. i suspect that they may get a
rude awakening with the TT, depending on who they thing the target market
is. if it is the public at large (i.e. non-commercial, like the ST),
they can probably get away with similar practices and succeed. if it is
commercial, then they would die unless they can sell a million of the
suckers in the first year. maybe they can. inertia will carry it then.
the up-and-comming conehead, beckmeyers, DC, et al will hack away to try 
and make a buck...

when u consider that u can run many mac AND ibm programs on the ST, it is
really a nice system, thanx to the hard work of so many talented programmers.
the last people atari (or any micro vendor) should want to rile is these
types of developers. they are the ones who made the ST a success (someone
once said: "software without hardware is an idea; hardware without software
is a space heater"). if developers want documented src for the os, atari
should just give it to them. they CAN get the ROM src themselves, but it
wastes SO much time which could be better spent developing s/w to sell more
h/w. if I were JT, i'd certainly do whatever i could to help the third
party s/w vendor, no matter how small, especially if it really didn't
cost me anything. if the src was in C, release it in C, not disassembled
code, please...

i used to complain mightily about various "quirks" in the ST os. now i
just accept them and work around them. part of the problem was my own
inexperience with microcomputer systems and the low-end s/w market in 
general. people who succeed in this market (or any market, for that 
matter) look past these problems and just DO IT. people who constantly
complain will never make it. if you made a bad decision, either live with
it or move on to something new...

part of the problem with the ST for more casual programmers (i.e. those
not relying on their efforts to pay the rent) is the lack of third party
books for the ST. i can find numerous books describing how to write
TSR's and device drivers for the PC. how interrupts work. what interrupts
are available. hardware interfacing. etc. NONE for the ST. you have to 
figure it out for yourself or have to be professionally trained in 68k
h/w design.

my idea of s/w (before the ST) was $50,000+ packages running
on crays and cybers (even blue boxes :^). big finite element and CFD
codes. NASA...GM...you know, BIG stuff...

for me, if i wanted to develop s/w for a unix box, i'd pick a dec/sun/next/
hp-apollo/etc box over an atari and sell to companies rather than hobbyists
(i would tend to favor linear address spaces *8^). only corporations and
universities buy them, and they tend not to pirate your work (since they
have far deeper pockets to go after if they did).

=BobR

BillR :^)
rosenkra@boston.cray.com