ZSR@PSUVM.BITNET (10/20/87)
Usually when I write an article like this I'm either frustrated, excited, or
ticked off. Actually, at the moment I'm all three of these put together.
This is a letter directed at Atari, and I plan on sending a copy of it to
Atari (probably c/o Ms. Clavaran; she has helped me before and I have the
utmost respect for her.) However, I'm not afraid to share my feelings on the
matter, and in order to see how others feel on the subject I'm posting this
little dissertation to the net.
Formal introductions aside, this is a letter regarding a recent posting to
the net: "More vaporware from Atari......" which I believe was penned
(keyed?) by Mr. Simon Poole. In a nutshell, the article described an
interview with Atari's head of research, Mr. Shiraz Shivji, concerning a
future product that includes a 68030 uP, Floating Pt. co-processor, SCSI,
VME Bus, networking hardware and other goodies.
The reason I'm upset is because Atari has been very much out of the marketplace
for quite a while and the market seems to be running away from them. In its
day the ST was an incredible machine from the "Power without the Price" stand-
point, and indeed I am *very* happy with my ST. Dain-Bramaged OS aside, the
ST has a lot of oomph and is getting me through college quite nicely. However,
it almost seems that Atari has deserted us, the owners of their products.
I realize that software upgrades and such take time, but how long has it
been since there was a substantial upgrade *MADE AVAILIBLE* to the public
(through the dealers)? The only one I can recall is the transition from
TOS on floppy to TOS in ROM. The new ROMs don't seem to be widely distributed
and the blitter is nowhere to be seen. TOS is a very slow environment to work
in, and I'm sure with more careful coding could be made faster. The folks at
Digital Research of course would probably have a large part in the speeding up
of the GEM routines, but even so, nothing seems to be getting accomplished.
Being a registered developer, I inquired about the GDOS and its availibility.
I was told that in order to obtain a copy I would have to register the specific
software that I was working on. Now, I'm currently not working full-time on a
software project (with school and all ;-)) but I do like to program in my
spare time. Atari is acting quite paranoid over a piece of software that
should have been part of the system to begin with. The whole essence of GEM,
in my understanding, was to have been a very general interface, supporting many
different types of printers, monitors, etc. and that GDOS was meant to
bind the GEM routines and I/O devices together. This has not happened, and
every company who puts out a product must write a separate driver for each of
the three bizillion printers on the market.
In Atari's defense, a new linker was released that VASTLY improves the link
times with the Acylon C compiler. They also came out with version 4.something
of the C compiler, which I do not have so I cannot comment on. Atari has
some very talented people who do read this newsgroup and reply (Mr. Harris,
Mr. Dyer, Mr. Faegan) and they do have some rather nifty hardware. The
philosophy of "Power without the Price" is, in my view, the business
methodology of the future, even though this motto has seemed to degenerate to
"Power without the Price or the Support to make it better."
Well, friends, we could go back and forth debating the above points until we
are blue in the face, but that is not *really* the reason I am writing this.
The preceeding few paragraphs have sort of set the scene and told us about
where we stand at the present (with the possible exception of the Megas; I
forgot about those and have heard that they are availible). Getting back to
the wonder machine that Mr. Shivji was discussing, let me say that I sincerely
hope Atari can develop and market it very soon. And while developing this
new system, I would like to offer some suggestions and food for thought to the
many talented people at Atari:
First of all, the decision to to with the 68030 uP is fantastic. From what
I've read about the 68030 and the entire 68000 series (see "Design Philoso-
phies Behind Motorola's 68000 Processors" (sic) in BYTE Magazine (rats, I
can't remember the exact issues, but was a three-part series in 1985) and
the more recent article "The RISC/CISC Melting Pot" by Thomas L. Johnson
in the April, 1987 issue of BYTE). People have been comparing the 68020
to the 80386, but they really should be comparing the 68030 to the '386.
Second, the Floating point processor is an absolute necessity when dealing with
numbers. The lack of hardware FP is one of the two crippling problems with the
ST. For any serious scientific or engineering work, FP support is a must.
The I/O is the other important fix to the ST dilemma. SCSI may not be God's
gift to I/O but it is fast and is supported (thanx to the Tan Toaster).
SLOTS--good heavens, that word from the mouth of Atari! Ever since the 800,
a major shortcoming of all Ataris has been lack of expandability. The Apple II
and the IBM PC in my opinion are not particularly good machines, but you can
DO stuff with them in the lab and in the classroom because they have SLOTS.
You can't be everything to everybody, so you let everybody MAKE you into
everything through the use of slots. Need A/D conversion? Plug in a board.
Need to control a robot or a milling machine? Plug in a board. Let the
Atari's horsepower control some mechanical horsepower, LITERALLY! Plus of
course the networking is a MUST for true classroom utility. I've seen both
stand-alone computers and networked computers in schools and the networked
systems have always been used harder and more completely (word of advice:
avoid CORVUS Omninet like the plague!)
UNIX (tm--this is the last "tm" that AT&T will get out of me for the rest of
this article) in my opinion is a nice OS. It may not be the fastest or newest
kid on the block, but it can be adapted to do almost anything (even real-time
stuff, if you have some imagination ;-)). It is expandable. It is well-
documented and well-supported. It can be made as friendly or as terse as is
needed. And as long as AT&T isn't greedy and doesn't demand megabucks for the
license, it is a pretty good OS for the new machine. Of course, Avante-Garde
can then allow us to emulate OS/2 with PS/2-Ditto ;-)
Basically what I'm saying to Atari is, "Look at the market. Look at what
people use; not at what the executives think is useful, but at what people
really USE. Look at the trends in the marketplace. Look at the FLOPS in the
marketplace (i.e. Closed MacIntoshes, heh, heh). Don't skimp on the software.
Do it well the first time. Offer support and upgrades (notice the word "free"
has not passed through my lips; Atari has to make a living too. As long as
it is "power without etc....") Give us some hot hardware and software that we
can mold and shape and call our very own (and it can be yours too if you have
an Atari computer and our new XYZ-ABC card...). Try not to cripple the machine
with inflexible hardware (such as the cartridge port that cannot be written to,
for want of another type of expansion slot). Don't be stingy with the tech
documentation because the third-party developers are there to help, not to
hurt sales. Don't charge $300 for incomplete, photocopied psuedo-documents
that make it difficult to develop stuff for the new system.
Things to do the same include maintaining a presence on this net, GEnie and
CompuServe, the "Power without the price" concept of marketing, the mouse
(I've really grown pretty fond of that little rodent), 3 1/2" disks (maybe
even the quad density drives), built-in RS232C and Centronics ports, and the
nifty styling of the machines.
Well, now I feel much better. I've gotten that off of my chest and now can
sleep at night. While I may not be a business major or marketing expert,
I am a USER and know what I (and those around me) want in a computer: the
moon and the stars. The only way that this will happen is with a solid,
expandable machine and a firm commitment to third-party and consumer support.
I sincerely hope that Atari succeeds in their new machines.
Christopher Johnson
CDJ at PSUECL
ZSR at PSUVM