|
Hi!
Thanks for all the kind letters and words of encouragement out there, TI folks
and others as well! In keeping with my series of articles on interviews, we
interview a professional TMS 9900 programmer, and a very good friend and a great
Brother too, David Ormand!
Hi
David! Please tell us a little about your self and your computer background!
And, your TI hardware setup? For those who may not know, David is a professional
9900 programmer-the CPU of
the TI 99/4A and 99/4. The Geneve uses a compatible CPU-the TMS 9995.
I have been a TIer for many years, having gotten the console kit less than a
year before TI dropped the Home Computer. On the week after TI's announcement,
the stores were dumping their stock
of TI equipment, so I got my mom to run down and we got a P-Box, 32K, RS232, and
TI Disk Controller at a fire-sale price of a few hundred dollars. Moving up
from the console/cassette recorder to a single-side, single-density floppy drive
was like discovering a new world!
Not too long thereafter, I convinced my parents to spring for Extended BASIC,
TI-Writer, and a Brothers electric typewriter that had an interface for a
computer. My mother's days of staying up
with me till after midnight typing term papers were over!
About the same time my Dad and I discovered the Southwest 99ers User Group,
which was a thriving club of thirty to fifty people. I got a 300-baud modem and
started logging into the club BBS,
the Cactus Patch. I discovered shareware. I joined in with some club
group-purchases of computer supplies, and participated in some group hardware
projects, the classic "mounting two half-heights"
and the "IBM power supply retrofit". Through the group, I purchased a copy of
Editor/Assembler, and started learning the guts of the TI-99/4A and how to get
it to do stuff that BASIC just isn't able to do.
After graduating from University of Arizona, my Dad insisted I quit sitting
around the house and get a job. I signed on with Hughes Aircraft Company, and
after five years as a manufacturing and test
engineer, the economy had some hard times. I was on the layoff list, but just
as I was due to go out the door, a friend I knew from church suggested I give a
resume to a department which was looking for software-capable engineers. I
cluelessly and fearlessly added my hobby experience on the TI-99/4A. God was
obviously behind all this, because it turns out the group who reviewed my resume
was working on TOW missile guidance set software, written in 9900 assembly, and
my hobby with 9900 assembly language got me a job with a great work group and
saved me from the unemployment line!
Since then, I've stayed with the 99/4A, and expanded it. I currently have in my
house a 99/4A rig with a CorComp double-side double-density controller running a
5.25" drive, a SCSI card running a 250MB internal drive and a 100MB ZIP, a 1MB
Horizon RAMdisk, and a 512K AEMS memory
expansion. This system is intended to pull off the former Southwest 99ers
shareware library, archive it on ZIP disk(s), and transfer it via shell scripts
and Magic File Manipulator to my Linux box, to be burned onto a CDROM. Since
this job seems to be a bit of a while off (too many projects, you know), my kids
use it for games and educational cartridges. Most of my work is done on my
Geneve, which also has a SCSI card running a 250MB internal drive, a CorComp
DSDD card running
two 5.25" drives, and a 3MB SNUG RAMdisk. I use it for my family budget (via MS
Multiplan), quick printout jobs using Myarc Writer, games (sorry, nothing much
more exciting than solitaire), and developing software in c99 and assembly. In
my "lab", along with some other loose consoles and P-Box cards, I have a 99/4A
with a CorComp DSDD card running a 5.25" and 3.5" drives and a 256K Horizon
RAMdisk, which I use for transferring files to 3.5" floppies, testing cards and
drives, and
eventually for networking with some 486es I've got out there, as work on
Ethernet and better RS232 devices for the 99/4A progresses.
Yes, I do have a Linux peecee, maybe to be supplemented with a PowerMac Linux
box sometime, an Atari Falcon inside the house, and an Atari STfm for playing
around with MIDI in my "lab". However, outside of an intent to keep up with the
"real" computing world via my Linux box, I have made a die-hard decision to
stick with the TI until the lights go out.
What exactly do you do for your company regarding the
9989? I mean what can be told!!!
The TOW (Tube-launched, Optically-tracked, Wire-guided) heavy anti-tank missile
has been in Army service since Vietnam. About twenty years ago, the Army
decided to upgrade their analog-circuitry Missile Guidance Set with a digital
computer. Hughes Aircraft contracted Texas Instruments to design and build the
MGS, and as our beloved company was wont to do, they used their own 9900
technology as the basis. There are two SBP9989 (radiation-hardened current
injection logic) microprocessors in an MGS, each with identical memory cards
with 48K of EPROM and 12K of
static RAM. One processor executes the optical tracker ("daysight") at 20Hz,
the guidance equations at 100Hz, and outputs the wire command signals via D/A
converters. The other processor controls the first-generation infrared imaging
tracker ("nightsight"), does image-searching algorithms, and passes a position
solution to the guidance processor via communications registers in CRU space.
The software for both processors is written in 9989 assembly language, which we
assemble and link using Texas Instruments development tools running under VAX
VMS. The basic Digital MGS system is used on TOW launchers on ground tripods,
HMMWVs, Bradley Fighting Vehicles, a few Marine vehicles, Cobra helicopters, and
several foreign fighting platforms, particularly in NATO countries.
Our group did the software for TOW 2B (fly-over, shoot-down), the Block II
maintenance upgrade, and several in-house mods to support enhancements and
demonstrations for new missile types and new platforms.
When you first found out about it, what were you able
to do with the 9900 clone, the SMJ68689? Any problems, solutions, etc?
Texas Instruments introduced the 68689 as a much faster upgrade to the military
users of the 9989. I think it was too little, too late, since almost all
development work any more is done in high-level languages like Ada, which is
just too big for little processors like the 9989. As a result, the Army decided
to do a "life-time" buy, and purchased all remaining stock of 9989 processors as
a strategic reserve. So not only are we not likely to ever use a 68689, but we
will see 9989s used in all legacy MGS derivatives ever to be produced in the
future. Actually, newer TOW launchers use processors that can handle Ada, like
the PowerPC and the Sparc.
When did you first encounter the TI 99/4A? (Or TI
99/4)
I have historically been a "I don't want it if I can't build it" type of guy. I
didn't want a calculator until I could build it myself - I almost succeeded with
this, although the Poly-Paks calculator kit was a chip, a keyboard, and an LED
display, so there really wasn't much to building it! I had the same
idea with computers, and was drawing circuit board layouts and writing base code
for the 6502 processor I had learned about in second-year digital logic at UA,
when the wars between Atari, Commodore, and TI really heated up. I decided to
postpone my do-it-yourself project and get a finished machine at the cheap $300
price. My Dad and I surfed the stores - Penneys, Sears, and
K-Mart - collecting literature, until I sold my Dad on the future of the 16-bit
99/4A (and color graphics, too!). It still seems amazing to me that our orphan
computers were sold as consumer goods off the shelves of retail stores.
What was your first program for the TI 99/4A?
I wrote a TI BASIC program that formatted a few characters into strategic block
patterns and used them to graph math equations. I was pretty proud of it,
although I'm sure my favorable recollection of it is colored by time. I saved
it on cassette tape, wish I could find it again...
How long were you involved with the SW99ers?
Not sure, 15 years perhaps. I got to be secretary a few times, and vice
president once or twice. The most fun we had was putting on Fest West in Tucson
a few times and Phoenix once, in cooperation with VAST (Valley of the Sun TI UG).
We met the TI personalities, saw the latest technology in action, and loaded up
on software and other goodies ordinarily available only via catalogues and
mail-order. We always tried to put on demonstrations at our meetings, and
support our members right up to the end. Lots of friends and good experiences.
Any major players in the TI Community have you met and
gotten to know about?
Fest West was good for this sort of thing. I saw the faces behind Comprodine
and Notung, met Bud Mills and Gary Bowser and Don O'Neil, and more luminaries in
the TI world than I can remember. I think I was most taken with Don O'Neil,
because I am attracted to hardware toys, and he was good at dreaming them up. I
got his keyboard adapter, more for the possibility of messing with the TI's
"operating system" in Console ROM than for the dubious privilege of using a
peecee keyboard.
Broken, now; I'm sticking with P-Box cards. Unless the latest thing I hear of
materializes - using Flash EEPROMs to replace the Console ROMs!
What did you think of the AMS/SuperAMS cards, and
please describe your work on them;programs, etc. (David was one of the initial
investors and supports) Any funny stories?
There have been several memory expansion schemes for the TI, including RAMBO,
SuperCart, and one brilliant-but-opinionated demonstrator at a Fest West who was
running his own system out of Horizon RAMdisk memory. Of all these, I think the
Asgard Memory System had the best shot at being
THE definitive common expansion memory architecture for the TI-99/4A. For
whatever reason, it didn't catch on, and by the time Asgard folded, turned the
engineering into the Public Domain, Jim Krych got the SW99ers to help, and we
actually were producing AEMS cards in quantities, the TI community was
contracting to the point where there were not enough developers to take
advantage of the new resources, or enough users to encourage developers to do
so. All the same, it was a real privilege to be part of the AEMS story, even
though the SW99ers did little more than provide funding, assembly, and sales
service to the project. Jim Krych did all the serious leg work.
Most of the fun came when we got the boards and the parts, and had group parties
to assemble and test the boards. There were maybe eight of us involved in this
phase, and we built 100 cards at Jack and BJ Mathis' home. Unfortunately, to
save money, we decided to use a very small card format, which lacks the keying
of other P-Box cards that ensures they can only be inserted correctly, so there
were a few instances of AEMS customers plugging their cards in backwards. One
of our own, Shawn Baron, plugged his in backwards, and blew out every card in
his P-Box. Another customer sent his back in for warranty repairs, but one look
at the scorch marks assured us of what happened, and the futility of repair (we
sent a replacement, and instructions to be careful to install it correctly).
You also wrote an emulator for the TI-an 1802, please
explain further.
I imagine many of your readers are familiar with, or even members of, the
Amateur Radio Relay League, and may know that the ARRL has put up satellites for
radio hams to bounce signals off of, or communicate with other hams at great
distances using satellite links. Turns out several of the big wheels of the
satellite project live in Tucson, and a fellow Hughes employee who was linked to
them knew about my passion for non-mainstream computers and processors. The
stabilization system of the AMSAT is controlled by a radiation-hardened CDP1802
microprocessor - there are other processors in the satellite to control the
communications and other systems of the satellite; if these get zapped, they can
be reset via remote command with no great harm done, but the
stabilization/orientation system is critical! Well, the 1802 is one of my
favorite "orphan" processors, and I was glad to help the AMSAT software
developers by creating an emulation of the 1802 and its I/O on the AMSAT. Of
course, being the sort of die-hard I am, I decided to write it in K&R-style c99
on the TI-99/4A, and port it to the peecee afterwards. Part of my reasoning
was, a few years before,
I got a friend to write an 1802 cross-assembler for the TI to support a project
of mine, so I already had the tool on the TI to feed test code to the emulator
on the TI. Got it working just fine on the TI, and had written it so the port
to the peecee was a piece-o-cake.
What is the best thing you have about the TI and the
computer community?
First off, the TI beats the pants off any other home computer of that era.
[Ooh, ooh, flame war!!]
Second, the 9900 processor family is just too unique and too much fun.
Memory-to-memory architecture, beautiful and largely orthogonal instruction set,
quick-n-easy CRU input/output. I deeply regret the modern trend that the CPU is
buried so deeply into the machine that it really isn't practical or even
feasible to write assembly code. I also regret the fact that Intel has won the
processor war on the desktop - the 68K family also has a beautiful instruction
set, which I have used both on Atari’s and industrial computers at
Hughes/Raytheon, and the PowerPC RISC machine has an interesting instruction
set, too. The trend is to use higher-level languages, from which all processors
look and act pretty much the same.
Third, the TI has so much I/O to play with. The cartridge port is a great place
to hack from. Best of all, it has a real card cage for expansion cards! I have
built several protoboards (just digital I/O),
some with RAM for writing device service routines. Fun, FUn, FUN!!
Fourth, a really fun community of people has grown up around the TI. Many are
gone, and not a few have left this life, but the die-hards remain, some via the
OLUG on Yahoo groups, others via comp.sys.ti, and many more I know remain,
isolated yet defiant, unconnected to the Internet or any surface mail contact.
The fact that new people are continually appearing on the OLUG is evidence to me
that the TI has real staying power, both from the attraction of its hardware,
the capability of even the early software, and the assistance and enthusiasm of
its gurus and die-hard practitioners.
On the flip side, any bad experiences?
Well, not personally, but like any other small community, we have our shares of
big egos and bombastic personalities. Many people have been badly burned by the
reaction of some loud voices in the TI community, and many have left because of
it. Maybe I have been more fortunate than most, or maybe I just take it for
what it is - hot air - but it isn't real healthy.
At the same time, it's much better than what I have personally experienced in
the Tucson Linux community, and what is reported, both by TIers and Atarians, as
what goes on in the Amiga and Macintosh communities.
What keeps you playing and hacking the TI after all
these years?
First, like I say, it's just so darn fun! Yes, I know the newer machines are
more capable: better graphics, faster communication, way, WAY more memory,
speed, and data storage. But there's something incomparable to seeing your
little 1970s machine process its beautiful assembly code
(or BASIC, or even c99 or FORTRAN-99!) that YOU wrote and do what you want it to
do!
Second, after it’s all said and done, it does most of what I want. Yeah, I've
got a Linux box with Abiword for letter-quality word processing, and obviously
it's going to do Internet communication that my TI will never be able to do.
But the TI has a good text-only word processor, very good for fast jobs, and I
know the convenient tricks to TI-Writer. Multiplan is perfectly adequate for
home budget use. Games and educational programs on the TI are frequently better
than anything on peecees or Macs, despite the comparatively primitive graphics.
And again, there is a real
kick to using a 20-year-old computer for common home computing tasks. Kind of
like seeing a Model-T Ford chittering down Broadway, with modern Escorts and
F-150s whipping past; it's GREAT to see from the outside, and you KNOW the
people inside are having an absolute BLAST!
What's your favorite software? Games, utilities, etc?
When I do my development work, I use My-Word on the Geneve, BA-Writer on the
99/4A, c99v4 for C-language and the standard TI Assembler for assembly code, and
the RAG linker and libraries. Someday I will get out TIC (the allegedly
ANSI-compatible C compiler for the Geneve), but c99 does everything I need.
On the Geneve, I work almost exclusively in GPL mode (TI-99/4A configured)
except for dropping back to MDOS 6.0 to move files between the SCSI hard drive
and the RAMdisks. When I am in TI mode, or on a 4A, I use the DM1000 disk
manager for file operations and the Birdwell Disk Utilities for sector hacking.
At some point, I want to get and try the new German utilities for file
operations on both standard floppies and SCSI drives for the 4A.
I'm not much of a games player, outside of a quick go at standard solitaire on
the Geneve now and then (have to be very light on the keys), but my favorites
continue to be Parsec (of course) and the Ray Kazmer Woodstock valentine Grog
game. I am also enjoying old things like A-Mazing and
Alpiner with my girls playing, not to mention helping with Meteor Multiplication
and various Forsman and Millikan titles. Soon, I want to have exposed them to
the joy of TI BASIC, and see what the next generation of TIers will do with it!
Any parting words for Retrogaming Times?
These are interesting years for computing and video gaming. Yes, the new
platforms have taken over the mainstream, and in most cases, considering the
demands of the new media, rightly so. At the same time, there is renewed
interest in the old platforms, not merely by old-timers looking
for nostalgia, but also newcomers discovering the old fun orphans for the first
time. So keep up the work of putting the old fun stuff before fresh new eyes!
|