I would venture a guess that most people who have owned a TI-99 for more than a couple of years have run across the name John Phillips before. He is a near legend in the TI-99/4A cartridge and assembly language programming community and can claim authorship, co-authorship or significant involvement in over a dozen cartridge programs produced for the 99/4A, not to mention numerous articles written about the inner workings of the 4A's architecture. Phillips will be 32 years old this year (1993) but he was only 21 when he was hired by Texas Instruments in 1982 right after graduating from Illinois State University. He started his career with TI in Dallas doing COBAL programming for business applications but it took him only 6 months to get a requested transfer to Lubbock where the "real" action was. John had purchased a 99/4 during his senior year in college and was already familiar with the Home Computer's architecture and he had wanted to program video games since purchasing his first cartridge, which was Munchman. Phillips didn't know TMS9900 assembly language but it didn't take him long to learn it.

His first project at Lubbock was Moonmine, followed by Hopper, which he co-authored with Michael Archulata. Hopper was followed by Word Radar, which he wrote in 2 weeks, for Developmental Learning Materials (DLM), the firm started by Bill Maxwell and Jerry Chaffin. After completing Word Radar TI sent Phillips to Japan where he met with several companies who were being recruited to write software for the 99/4A. Following his return from Japan he became involved in almost every piece of software that was slated for production or that was actually produced for the 99/4A. When TI announced the end of the Home Computer Division Phillips was offered several incentives to stay at TI but turned them all down because none involved work with the 99/4A. Instead, he and fellow employee Michael Archuleta went to work for DLM, which had continued to work on products for the TI-99/4A even though it was no longer being produced. In December 1983 John Phillips announced to the TI Community that he was available to any User Group for seminars, demonstrations and question and answer sessions related to the TI-99/4A. He would travel to virtually any location if the User Group would pay round trip airfare from Dallas, Texas plus lodging? While he could only make himself available on weekends, it was a pretty generous offer.

Both Phillips and Archuleta eventually left DLM (probably because the work there dried up too) and started their own firm in February 1984 called Video Magic. Video Magic also came to an end in too short a time, I suspect because it was becoming painfully obvious that one could not make a living trying to write software for the 99/4A.

At Texas Instruments Michael Archuleta was responsible for the 99/4A Technical Hotline and for 99/4A software quality assurance. Phillips was a third-party software development consultant and programmer in the education/entertainment section of the Consumer Products Division. Both men would get together again in 1986 to collaborate on the 4A FLYER game cartridge that was commissioned by Triton Products. To date, that is the last time we've heard from the John Phillips/Michael Archulata team.

Archuleta and Phillips were involved in, or responsible for such TI-99 favorites as:

ANGLER DANGLER - Phillips worked on this project as the debugger of the final code, but the project never reached completion before the bailout so Angler Dangler was never officially released. It does exist in GRAM file format however, so it probably was not too far from being a real product when someone at TI made the decision to pull the plug. If you look at the October 23, 1983 IUG price list you will see Angler Dangler listed as being available.

BEYOND PARSEC - This cartridge, which Bill Moseid's DataBiotics firm released for the 99/4A during the third quarter of 1988, started life in early 1984 as one of two game cartridges John Phillips was writing for CorComp's new CCI-99/64 (aka Phoenix) computer. The other game was Star Wars. Both efforts came to a screaching halt however, when TI objected to the use of the Parsec name, and George Lucas' company apparently objected to the use of the trademarked Star Wars name. The Star Wars code must have actually been finished at the time though, because I have the game on disk as a GPL file. It was ultimately renamed Star Trap and released in cartridge form by Exceltec in 1985 and then by DataBiotics during the third quarter of 1988.

BEYOND SPACE - This is a John Phillips creation that was completed in May 1984, but not released until the first quarter of 1985 when Exceltec/Sunware marketed it. It was picked up by Unisource Electronics for their catalog/encyclopedia but pretty much floundered and then just disappeared. It has never resurfaced since both Unisource and Sunware went out of business in 1986.

The game involved two players with each having a ship of equal firing power. The area in space where the two ships confront each other is littered with asteroids which may be moved by firing the ship's laser. The object of the game was to push asteriods into your opponent's space ship to crush and destroy it. The only review I've ever seen written on the program claimed that its speed was too fast to play the game very long, so that may be why it has slipped into oblivion?

BURGERTIME - Phillips provided the final debugging for Burgertime.

D STATION - This John Phillips creation has the distinction of being the only program ever released by the International 99/4 User Group on the Romox ECPC cartridge. You may recall that during the fourth quarter of 1983, Charles LaFara promised "a library" of programs from the IUG on the Romox ECPC (Edge Connector Programmable Cartridge). D Station was just the first, but it also turned out to be the last.

When the IUG ECPC library failed Exceltec (aka Sunware) picked up the program and marketed it for a short time in 1985. Triton finally introduced D Station in their Fall 1988 catalog along with a brand new D Station II game, also written by John Phillips.

D STATION II - See D Station.

FACEMAKER - Phillips collaborated with Intersoft's Jerry Spacek on this project. Spacek you may recall wrote Defend the Cities, which was the first commercial Mini-Memory assembly language game ever written. In the Facemaker project Spacek translated Spinnaker's source code to TMS9900 assembly language and Phillips ported it to cartridge format.

HOPPER - Michael Archuleta and John Phillips co-wrote Hopper, which was the only cartridge developed entirely on the TI-99/4A Home Computer, using the Editor/Assembler cartridge for all of the programming. All of the other TI-99 cartridge software programs were developed on a TI Mini, not the 99/4 or 4A. [All of the programs for the Mini Memory cartridge were programmed exclusively using the 99/4A and Editor/Assembler cartridge. As noted at the beginning of this article, all 5 of the titles developed by Sofmachine were also programmed using only the 99/4A. I suspect that many of the third-party games were also programmed in this manner, but I can't say for sure. ]

JAWBREAKER II - Phillips converted the original Sierra On-Line source code to TI-99/4A code.

MINI MEMORY'S LINE-BY-LINE ASSEMBLER - Phillips claims responsibility for its development, but I am not sure exactly what that means. [I developed the Line-By-Line Assembler excusively. John Phillips was not even working in the Home Computer division at that time. And I can't claim responsibilty for writing all of the code, only for porting it from the 990/189 University Board single board computer. Also, I wrote the Lines program. I am not aware of ANY contributions which John made to the programs in the Mini Memory. ]

MOONMINE - Programmed by John Phillips from a design by Bob Hendren. You may remember that Hendren was also the project engineer behind Parsec and the person who recruited Aubree Anderson to do the voice for the Parsec game. [Bob Hendren had ABSOLUTELY NOTHING to do with the development of Parsec in regards to content, playability or technical direction. With respect to his role in the recruitment of Aubree Anderson, I really don't know about that. Parsec was strictly a collaboration by Jim Dramis and myself. Parsec was not directed or defined in any way by management or anyone else. We were merely instructed to "..get together and see what you can come up with...". Although we received much input from our coworkers as they played with Parsec during the development process, almost all of Parsec was what came from our imaginations. I think I can speak for Jim when I say we are still pleased with our efforts more than 10 years later. We both hope that everyone who has played Parsec has enjoyed it as much as we enjoyed both writing AND playing it. ]

PETER PAN'S SPACE ODYSSEY - Phillips and Archuleta collaborated on this program while employed at DLM. It was never officially released but is available as a GRAM file that can be run from P-Gram, Gramulator or the GramKracker.

SLYMOIDS - Slymoids was written by James R. Von Ehr II. The cartridge conversion was accomplished by John Phillips.

STAR TRAP - See Beyond Parsec.

SUPER DEMON ATTACK - Phillips worked on this project, but I have no information on the specific contributions he made to its completion other than possible debugging of the final code. I do know that he actually worked on Demon Attack, not Super Demon Attack, but they are probably the same project with the actual marketed product just having a slightly different name.

THE GREAT WORD RACE- John Phillips authored.

TREASURE ISLAND - Phillips provided the final debugging for this game cartridge, which had apparently become stalled by a bug that no one could find.

WORD RADAR - John Phillips authored.

-------------------------Paul Urbanus Clarification 10/13/1995------------------------

I would like to clarify and correct some of the statements made in the original posting of the below article. To help substantiate my comments, I will start with a history of my relationship with the TI 99/4A Home Computer.

I was hired by TI in December of 1979 as a coop student, and spent six months working for the Home Computer Divison of TI in Lubbock. During this time, I did such diverse things as gather statistics on the distribution of data in the various GROM chips and program an HP test station to verify that the RF modulators were meeting FCC and TI specifications. In my spare time, I purchased a surplus IBM terminal keyboard (a really nice one) and interfaced it to the 99/4. This was the first REAL keyboard for the 99/4. I even added logic to do auto-repeat and mapped the IBM cursor keys to function correctly. My coop term expired, the fun ended, and I went back to school (New Mexico State University) for a year.

I returned to TI in Lubbock in the summer of 1991 to serve a second stint as a coop student (and earn some money!). In my absence, the TI99/4 had undergone puberty and blossomed into the TI99/4A. In addition to the new keyboard, there was also a new video chip. The TMS9918 had been replaced by the TMS9918A. At this point several things were happening, and the confluence thrust me into a rather unique position at TI.

My first assignment was to perform some testing of the new video chip and plot a chart of chip operation versus supply voltage and temperature. While waiting for the temperature chamber to stabilize during these tests, I was reading the detailed chip specification and came to a startling discovery - there was a new graphics mode in this chip which would allow neat new applications. At the same time, the Editor/Assembler (E/A) cartridge was in the early stages of alpha (internal) testing. I used the E/A cartridge to play around with the new graphics mode (Graphics Mode II). One of the first programs I wrote was a simple line-drawing program which I called "Lines". This is the same program which was bundled with the Mini Memory module.

After I wrote the Lines program, management moved me from the hardware to the R&D group and suggested that I collaborate with Jim Dramis on a new game. I thought this was better than sex (ok, I WAS kinda naive) - getting paid to write a video game. Just for reference, Jim had written some of the best TI games available at that point - Car Wars and MunchMan. We quickly agreed that we wanted to write a space game and we wanted to have smooth horizontal scrolling to give the illusion of flying over the surface of a planet. As some of you may know, there is NO hardware support for scrolling the screen on a pixel basis in the 99/4A video chip. After lot's of pondering, I hit upon the solution - copy the inner loop of the scroll code into the fast 16-bit RAM of the 99/4A. Since this code is responsible for 80% of the execution time of the scroll loop, substantial speed gains were made by moving the loop to fast RAM. In today's world of 486s and Pentiums, this RAM would be referred to as cache RAM. I then handed this code off to Jim so he could incorporate it into the game.

The next thing I wanted to do for the game was to come up with some really neat sound effects. Since the sound chip on the /4A was only capable of generating square waves, I wanted to use the speech chip. The speech chip operates by using a model of the human vocal tract, and I reasoned that if people could make really strange noises, then so could the speech module. After studying the speech chip specification, I made an important discovery: the speech chip didn't need new data very often (it sure helps to understand hardware when writing software). This fact could be combined with one of the new features in the 99/4A software architecture - the User video interrupt. The net effect of this combination was that the speech chip could be used while the game was going on. When I went to the software folks with my discovery, they told me that "you coun't do that". Only after I showed them did they believe.

I created the asteroids in Parsec in TI Logo. I wrote a small Logo program to animate them, and iterated the shapes until they were satisfactory to me. Then I wrote an assembly program to convert the asteroid bitmap from the binary TI Logo data file to ASCII data statements for use with the 9900 assembler.

All of the above programming was done on the 99/4A using the Editor/Assembler package. EVERYTHING I wrote for the 99/4A was written using the Editor/Assembler cartridge. I liked it much better this way, because I could work at home, and I could fix the /4A system if it went down, unlike the 990 minicomputers.

As Jim continued to progress with Parsec (we brainstormed on ideas, but he did most of the game flow implementation), the Mini Memory cartridge was developed. However, there was no software available to make it do anything useful. So I suggested that this would be a great tool for letting people experiment with assembly language without having to have any peripherals other than a cassette recorder. The Line-by-Line Assembler was a derivative of the code used in a TI single board computer which had been developed for microprocessor courses at the university level. This single board computer was called the University Board (model no. 990/189). When I returned to school after my first coop session, I had borrowed one of these from TI and it was an excellent learning tool for me so I assumed that a similar capability on the /4A would also be good.

We were able to get the source code for the assembler from another TI group. All the I/O routines expected a dumb terminal, and so they had to be converted for use with the /4A keyboard and screen. I also added a routine to dump the symbol table. In retrospect, the code could have been a lot cleaner and more compact, but I can probably say that about any program I write today after I have finished it. We decided to include the Lines program as an example of how to program the new video chip, as well as instant gratification for Mini Memory customers.

My final task was in this coop session was to go out to La Jolla, California and work with Control Data Corporation and educate and support them in their efforts to port the Plato series of computer-based courseware to the 99/4A. I spent about a month out there, and in that time I wrote the graphics and disk I/O package for the Plato interpreter. A byproduct of this work was an intermediate tool, DISKO, which was used for debugging the disk I/O package. I understand this program eventually made it into the public domain. For those of you familiar with this tool, there is a whimisical menu choice, "Resign/Go to Black's Beach". Black's Beach is a nude beach in La Jolla :)

I returned to school in the fall of 1992, but only lasted for one semester. At that time I joined with my Parsec partner, Jim Dramis and the author of TI Invaders, Garth Dollahite, along with two business types and we formed a company called SofMachine. Our charter was to author, produce and market game cartridges for the TI 99/4A. Kind of like the TI version of Activision. While Sofmachine was in existence, we wrote three games of our own and converted two games for Atarisoft. The games we wrote during that time were:

Because our business partners were unsuccessfull at securing the required venture capital funding, combined with TI's exit from the home computer market, we were unable to manufacture and market our (Sofmachine's) three games. However, due to a sequence of events beyond our control the Sofmachine games were pirated and eventually freely exchange around the TI 99/4A community. A valuable lesson was learned: NEVER trust anyone with your own livelihood. Lesson number two: Don't believe what a "business" guy tells you just because they're the business guys and you're the technical guys, ESPECIALLY if it goes against your gut instincts.

A number of years later, the Sofmachine games were released in a cartridge form by Databiotics under license by Sofmachine.

All games programmed by Sofmachine used the TI 99/4A as the development platform, along with the Editor/Assembler cartridge. Two of us programmers purchased a Myarc 10 MByte hard drive for $1800 EACH! I just sold it about 6 months ago for $100. OUCH!

Wow! I intended for this to be a brief history for the purpose of lending credence to my following comments on the John Phillips article. Sorry for the long windedness, but the recollection of these times is good and I almost couldn't stop typing.

Please see below for my rebuttal to specific points. Thanks and regards, Paul Urbanus the "Urbite" in Parsec urb@onramp.net