Talk:TI-99/4A

Content merge
I merged content from TI99/4A. Does it still read well? --Timc 14:56, 12 Aug 2004 (UTC)

Other opinions
This is a nice article, but it isn't all that honest. Don't get me wrong, I have a certain fondness for the TI, and I've written an emulator for it... but it was never a usable or competitive computer.

The TI-99 wasn't "technologically competitive", especially after the IBM PC was released. Even though it had a 16-bit CPU, its severely crippled video display capabilities, lack of directly addressable RAM, and restricted programming capabilities made it a poor choice for any serious applications. Even a 6502-based machine at 1mhz could run rings around it. — 12.103.251.203 &#x20;07:56, 4 September 2005 (UTC) — continues after insertion below


 * Keep in mind that the TI was never meant to compete in the same market as the IBM PC, predates it by two years, and cost 1/6th the price of a PC! The TI-99/4 was originally designed at a time when it was assumed that the general public would want a computer for doing their taxes and playing video games; Texas Instruments (and many others) were caught by surprise when the home computer revolution spawned computer hobbyists who actually wanted to learn how to program in Assembly. :) — Preceding unsigned comment added by 216.138.194.68 (talk) 06:03, 5 February 2006 (UTC)


 * There exists information that states the TI-99 was original to be a game machine, a project between TI and Milton Bradley (which made a lot of the early games). As a game console, it would not need a lot of RAM... all the games would have been cartridge-based. The video chip (TMS9918, and later the TMS9918A) would have been perfect for games. Ti99 forever (talk) 11:54, 7 March 2014 (UTC)

Its graphics capabilities were, frankly, horrible! Yes, it had a lot of sprites... and that's about all. Programs couldn't really do much graphically, thanks to the unavoidably slow VDP access. (It could easily take 30-40 microseconds to update one pixel, while on most of the then-current 8-bit machines screen contents could be updated at CPU clock rates during HBI/VBIs.) Lack of rational bitmapped graphics modes didn't help either. Scrolling was a serious challenge.

Lack of useful collision detection between sprites made them astoundingly useless for games, as all the VDP could tell you was that two or more sprites overlapped--not which ones! Programs rarely used most of the sprites, because only four of them could appear on any given scanline (the rest were blanked); witness Pole Position's awful monochromatic "enemy cars". There's no way it competed graphically with either the C64 or the Atari, at least not in terms of natively usable horsepower. — 12.103.251.203 &#x20;07:56, 4 September 2005 (UTC) — continues after insertion below


 * The C64 was introduced in 1982, the Atari 400/800 also in 1982. Remember that the TI's closest contemporary competition was the 1977 Apple II which was designed (and priced!) as a business computer, the equally expensive business-oriented PET, and the VIC-20 which can only be dismissed as lame (with the aluminized cardboard RF shields in the VICtim-20/C-64 versus the nickel-plated steel in the TI, your analogy is like comparing a Hyundai with an automobile). The /4 was introduced in 1979 and was extremely competitive in its day; the /4A was an enhanced but mostly software-compatible replacement which didn't sacrifice compatibility (after all, users weren't interested in understanding why their new computer wouldn't run the software already available for it).


 * Furthermore, Pole Position is a bad example. It was released by Atarisoft, based on reverse engineering of the machine rather than by actual formal documentation from TI.


 * I flaw TI on a few things, the biggest of them being the lack of schematics and other technical documentation being provided by TI. They didn't know any better at the time and wanted to corner the market. I flaw TI for not extending the 16 bit bus out to the expansion port so that multiplexed and 16 bit RAM expansions (and other 16-wide peripherals) could be optionally available. (I think I'd call one "Basic RAM Expansion" and the other "Turbo RAM Expansion", but this is with over 25 years of hindsight in the personal/home computer marketplace under my belt.) I flaw TI on the double-interpreted BASIC (BASIC -> GPL -> machine code). I flaw TI for not having added a video input jack and genlock into the machine, since the VDP was capable of external sync and the VCR and home video cameras were just starting out and it would have been an exclusive to be able to corner the market on titling your own home movies (ie. a Video Toaster 6 years ahead of its time). I flaw TI for not having a decent (bigger than 48 key) keyboard - surely at least some prospective home computer users had taken touch typing courses. And I flaw TI for not making all graphics modes available directly from BASIC interpreter built into the machine. But hey, hindsight is 20/20; at the time, no one cared - what still blows my mind is that TI actually stuffed a 16-bit processor in there, they must have had an 8 bit processor which would have been cheaper to use!


 * Finally, as an electrical engineer with an understanding of marketing, what really blows me is TI's advertising. TI was good at selling to engineers - and still is - all I want is specifications. While TI was showing a picture of a TI-99/4A on the box, Commodore's box was a big and colorful thing showing "Commodore VIC-20 - With color and music!" (nb. less colors, less resolution, and less voices - to say nothing of RAM or CPU horsepower which 1981 K-Mart shoppers didn't understand). Amazingly enough, TI didn't even mention the 16 bit processor (I can see the ad line: "Double the bits, double the power!" Let alone "Two cassette ports built in! Two Joystick ports built in!", etc.) on the boxes or in *any* North American advertising that I saw as a TI user and afficionado during the home computer era.


 * Summarizing TI's faults: lack of 16 bit bus extended to the expansion port allowing high-speed expansion; lack of a video input jack which would have cost an LM1881, one capacitor, and two resistors; lack of any idea how to sell *anything* (let alone computers) to the unwashed, ignorant and uncircumcised masses; and lack of technical documentation with the understanding that this would promote platform development and therefore sales. But hey, it was 1979. — Preceding unsigned comment added by 216.138.194.68 (talk) 06:03, 5 February 2006 (UTC)


 * You can speed up the memory access to the video processor by performing access during the access window (same as a VBI), or by toggling certain functions off (setting the BLANK bit, which hides the screen... turning off VDP interrupts, sound processing interrupt, etc). Finding which sprites overlapped was not hard -- just read the x-y coordinates into RAM and compare.


 * As for the 16-bit processor... it was originally going to house an 8-bit TMS9985 chip, the predecessor of the TMS9995 used in the unreleased TI-99/8 and in the Myarc Geneve 9640. But the TMS9985 was a failure, and TI had to stick the older TMS9900 in which required static RAM chips (the 9985 had internal RAM), and then they had to add logic to the motherboard (multiplexer) to convert between 16-bit and 8-bit. Bad part is, being the 9900 came out in 1976 -- even though it was 16-bit -- it performed those operations as two 8-bit cycles! Ti99 forever (talk) 11:54, 7 March 2014 (UTC)

But the computer's true "Achilles' heel" was that TI didn't want anyone writing machine-language programs for it! TI only officially authorized GROM-based programs, which ran very, very slowly (both because they only supported serialized data access and because they were interpreted). Its built-in BASIC was one of the worst on the market, and the fact that it was also GROM-based didn't help. — 12.103.251.203 &#x20;07:56, 4 September 2005 (UTC) — continues after insertion below
 * Agreed. You could almost hear the tic-tic-tic of that thing incrementing the address register when simply running { 10 CALL CLEAR :: 20 A=A+1 :: 30 PRINT A :: 40 GOTO 20 }. Again defending TI, I think the focus was more about user friendliness (big title graphics and menu, etc.) because no one thought that speed or power were actual selling features. (Killer is the VIC-20, new in box in my closet, never used: "VIC-20: The FRIENDLY computer" would just disply "4021 BYTES READY" if it were powered up.) — Preceding unsigned comment added by 216.138.194.68 (talk) 06:03, 5 February 2006 (UTC)


 * GPL (GROM-based programs) actually run very fast. About the speed of other machine's BASICs. The GPL interpreter is written in assembly, after all. BASIC was slow because it was written in GPL (double interpretation), AND programs resided in VDP RAM, which is much slower than CPU RAM. Ti99 forever (talk) 11:54, 7 March 2014 (UTC)

TI Logo, one of the more memorable software products for the machine, was simultaneously a pinnacle of absurdity. Because of the lack of memory it had to place restrictions on what could be drawn on screen. If the on-screen display became too complex (more than 256 unique tiles) you'd get an "out of ink" error. (Remember, the computer had no native RAM to speak of; video memory had to be shared with the Logo program. If it had used a true bitmapped display it wouldn't have had enough memory left for any significant programs. TI Logo-II got around this by requiring a PE box plus memory expansion.) — 12.103.251.203 &#x20;07:56, 4 September 2005 (UTC) — continues after insertion below


 * First off, the RAM expansion was available without buying the PEB ("PE Box"), it was a popular standalone option. — 216.138.194.68 &#x20;06:03, 5 February 2006 (UTC) — continues after insertion below


 * "Out of ink error" occurred because bitmapped mode was not used. Logo could be used with the 99/4 and the 99/4A, the 99/4 not having a bitmap mode. Plus, bitmap mode consumes *a lot* of memory. Logo requires *a lot* of memory. So in Turtle mode, Logo was actually redefining characters, and there are only 256 characters in the character set. And both Logo and Logo II required the 32k memory expansion. Ti99 forever (talk) 11:54, 7 March 2014 (UTC)


 * TI Logo was neither memorable nor popular, though I don't think it had anything to do with that - I don't think Logo was popular at all outside of schools. TI did release *many* cartridges which allowed full bitmaps - Parsec, Alpiner, etc. - which included static RAM to alleviate the VDU RAM issue. — Preceding unsigned comment added by 216.138.194.68 (talk) 06:03, 5 February 2006 (UTC)


 * The above statemnt is not correct. Neither Parsec nor Alpiner had static RAM.  Both had GROMs and ROMs, but no RAM.  The only cartridge with static RAM was MiniMemory. — Preceding unsigned comment added by 64.221.222.142 (talk) 15:39, 11 September 2006 (UTC)

Frankly, it's truly impressive to see just how slowly a 16-bit CPU can be made to run. The 9900 had some great ideas (if nothing else, register windows!) but the 99/4A hardware or software never took advantage of any of them. — 12.103.251.203 &#x20;07:56, 4 September 2005 (UTC) — continues after insertion below


 * The machine was actually quite fast when running *anything* with real CPU RAM, despite the 8/16 clusterfsck; certainly competitive with anything in its day and even extending beyond (C-64, etc.). I highly agree, though, neither hardware nor software did anything with it. I can understand the lack of CPU RAM in the console (even a few k was very expensive in a day when it was assumed that most users wouldn't care about a computer's speed). But okay, so let's make the expansion port 8 bits wider - that's a cheap card-edge connector and a few more traces on the PC board; so far, we're not breaking the bank and we've got future possibilities - why the hell didn't they do it? Dunno. Maybe even sell a 16-bit 32k RAM expansion at cost to promote the machine?


 * GROM itself makes sense for most users (who will never buy expansion peripherals, I guess they figured), but I really think the BASIC interpreter should have been in ROM.


 * (Note to non-TI experts: GROM is "Graphics" ROM, a self-incrementing ROM developed by TI, and which was usually relatively small and cheap at the expense of speed. A common hack with the 99/4A is to desolder the BASIC GROM and replace it with the GROM from the Disk Manager 2 cartridge, effectively building DM2 into the console.)


 * And finally, the register windows. Yeah, registers in RAM were great (a TMS9900 feature). I've seen many upon many TI programs which took advantage of the near-instant context switches, allowing the machine to multitask based on the human-imperceptible vertical interrupt coming from the TMS9918. There's lots of TI-99/4 and TI-99/4A software using this feature. What year did DesqView come out? (I dunno, I was using my TI-99/4A until 1990 when I bought an Amiga, and that until 2000 when Linux was finally viable as an operating system for the x86 platform.) — Preceding unsigned comment added by 216.138.194.68 (talk) 06:13, 5 February 2006 (UTC)

The final stake in its heart was the lack of affordable expansion capabilities, as most of the interesting expansions required the absurdly expensive PE box (basically a card cage with a power supply... only $900 or so). — 12.103.251.203 &#x20;07:56, 4 September 2005 (UTC) — continues after insertion below


 * 32k RAM expansion, disk controller (3 drives without the sleazy Commodore I-can't-believe-it's-not-a-cassette serial interface speed issues), RS-232 port (2 serial, one parallel), and the most popular speech synthesizer were all available without using the PEB. Admittedly, the PEB was insanely expensive, but it was also insanely overbuilt; I suspect TI Micro farmed out the design to TI Mini... or maybe even TI Military. — Preceding unsigned comment added by 216.138.194.68 (talk) 06:03, 5 February 2006 (UTC)


 * One should realize that the Peripheral Expansion System was not always expensive. In late 1982, TI had a promotional offer. Buy an expansion system with your choice of two from 32K memory expansion, RS232 card or Disk controller+1 drive for $400. And the choice of TI-Writer word processor or Microsoft Multiplan. The PES (as it was named by TI) had only been introduced a few months before. This was no "fire sale". Ti99 forever (talk) 11:54, 7 March 2014 (UTC)

The 99/4a was intended to be very cheap to produce hardware-wise, hence the serialized ROMS, VDP and such. But it never reached that goal for a couple of reasons: the promised 8-bit version of the 9900 CPU didn't happen so they had to release it with an expensive 16-bit version, and there was very little interest in their other serialized hardware chips so they were never that economical to produce. — Preceding unsigned comment added by 12.103.251.203 (talk) 07:56, 4 September 2005 (UTC)


 * And yet again, TI was well ahead of their time. USB. Firewire. Serial ATA. Texas Instruments HexBUS interface. The Texas Instruments TI-99/4A. — Preceding unsigned comment added by 216.138.194.68 (talk) 06:03, 5 February 2006 (UTC)
 * And there's still nothing like the flicker of the PEB's access lights then the clunk-clunk-clunk of the old Shugart reading 90k off what now seems to be a ridiculously oversized diskette on hardware which is now as old as many of my coworkers. To t1heir credit, the PEB, TI Color Monitor, the console, and all other TI-built home computer hardware I've ever owned have never suffered a hardware failure. I still have a console, "Wired Remote Controllers" (joysticks), speech sythesizer and cartridge drawer in my living room, and another system (complete with PEB, CorComp disk controller and Horizon RAMdisk) beside the machine I'm currently using. (It's part of my homage to Generation X.) — Preceding unsigned comment added by 216.138.194.68 (talk) 06:13, 5 February 2006 (UTC)
 * Please summarize your point, and please sign ur comments. --Soumyasch 10:42, 24 February 2006 (UTC)

TI and the Education Market
When I went shopping for a home computer in 1983, the guy at the TI store asked me if I were a teacher. I admitted that I'd taught 7th-9th grades for a year and a half three years ago, and presently taught an occasional college course in the evening. Well, that was enough to get me a small discount on a 99/4A. I think I paid something like $200 plus tax for the basic machine.

I regretted that decision pretty quickly when I learned that ANY expansion of my system was going to cost a thousand bucks. Sure you can get a disk drive ($250). . . and you'll need a disk controller ($300) and the peripheral expansion box. (a million dollars) More memory? Same deal. Printer? Take out a loan. (TI was trying to bury the stand-alone expansion items and didn't talk about these.) I bought a Commodore 64 for serious work, and it cost me a lot less.


 * In late 1982, an expansion system package was available for $400. Your choice of two of the following: 32k memory card, disk controller (with disk drive) or RS232 card. And you got to pick as a bonus either the TI Writer word processor or Microsoft Multiplan. As the RS232 card had both serial and parallel, and these were universal interfaces, you didn't have to buy a TI printer. Ti99 forever (talk) 12:02, 7 March 2014 (UTC)

However, the TI-99/4A was well adapted (with Extended Basic and the Speech Synthesizer) for writing simple games and learning tools for my young son to use. The kid knew the alphabet before he was a year old due to playing with a simple thing I'd written to display a letter enlarged and speak its name. The cartridges were easily swapped in and out, so he could play different games (many of these learning-oriented) without getting bored.

TI continued to support "teachers," however nominal. When my old console had problems several years later, they took the battered, dusty old model in trade for a brand new one at no additional cost, due to some clause in the original deal.WHPratt (talk) 14:13, 4 March 2009 (UTC)


 * TI had a very nice policy on repairs. Up to the end and beyond (1983) you could send in your unit, get a replacement, and at most pay $50. Ti99 forever (talk) 12:02, 7 March 2014 (UTC)

wot no picture?
Anyone fancy adding a picture to this article? --Rebroad 16:44, 1 November 2005 (UTC)

Also, I want to add the hardware palette file relevant to this computer. WikiPro1981X (talk) 14:10, 2 February 2010 (UTC)

games section?
What about having a section in the article on the TI's (most popular) games, such as Parsec, TI Invaders, Munch-man (Munch Man or Munchman?), and Car Wars. Those four stand out to me. 24.222.121.193 01:17, 22 April 2006 (UTC)

Arthurvasey (talk) 23:55, 19 May 2009 (UTC)

We bought a TI 99/4A out of the local paper once - it came with cartridges for Invaders, Munch Man and Parsec - it also came with some demo cassettes with games like Space Cities - they were written in BASIC, so you could rewrite the programs - my brother managed to add a line or two where you could input how many lives you wanted - he put about a million in and it came up as a funny number - something like "115E+20" or something weird like that.

There was a game involving tanks - took AGES to load.

There were also two programmes which looked like promising games - but "Firework Night" was nothing more than a recreation of a rocket firework being launched into the air on a loop and "Merry Christmas" was an intricately designed program that, in large letters, flashed up "MERRY XMAS" on the screen on a loop.

My brother tried to design a program for Yahtzee - but the repeated use of GOTO and GOSUB in the program caused it to keep saying "MEMORY FULL".


 * You probably had some recursion going on, and ran out of stack space. Ti99 forever (talk) 12:08, 7 March 2014 (UTC)

My brother's most ambitious project was to write a program for the Channel 4 game Countdown - it took him several attempts to get it to run properly - making it so you just pressed V for vowel and C for consonant, without having to press ENTER, getting the numbers game to automatically generate two from the top and one from the others (that's the best way of getting to crack the numbers, we found), creating a clock that counted down 30 seconds (he couldn't reproduce the proper theme, the memory being limited and very few tones available) ...


 * The TMS9919 sound chip has three tone generators that range from 110 HZ up to 44,733 HZ. I'd say "that's a lot of tones". And there was a noise generator that had 6 fixed noises (3 "periodic" buzzer-type, 3 "white noise"), and one of each type that was programmed by the frequency given to the third tone generator. Ti99 forever (talk) 12:08, 7 March 2014 (UTC)

He even managed to create different coloured screens for each round - in LIST mode, it was black on a cyan background - if you didn't have a colour change in your program, the background defaulted to green in RUN mode.

For the reason detailed above, it was difficult to do any of the following with the Countdown program:

Whilst he managed to get vowels in the Vowels and consonants in the Consonants section (so that, when you asked for a vowel, you didn't get a Q), some letters were often repeated;

Same with the numbers - there were often two 25s or two 75s in the top row (he had managed to just draw on the numbers 1 to 10 for the bottom rows, so you wouldn't get a 15 come out);

It was not possible to teach it maths or English - so you couldn't get it to make its own words up or solve the numbers - you had to do that on pen and paper.

He came unstuck, however, when he tried to add the Conundrum round - it worked fine for all words beginning with A - he picked out some words like ABAMPERES and ABACTINAL (we didn't have a good dictionary at the time) - but then, the MEMORY FULL kicked in.

Somebody we knew with computer knowledge suggested we used the OPEN command and DATA statements - but that didn't work, either.

Google Images
The external link to search for images of TI99/4A is labelled as Google images. But the actual images come from somewhere else. Google merely indexes them. So the link title should not be Google images. Actually, link to search for a particular item should not be kept. The user can do the search himself if she wants to. --Soumyasch 10:40, 24 February 2006 (UTC)

I can add a picture of an original TI-99/4A retail box, if anyone sees any use in having that on the wiki. Alex W 21:08, 13 March 2006 (UTC)

-- Go for it! The TI-99/4A's retail box was another horrible marketing blunder! While Commodore was selling the VICtim-20 in boxes bragging "The Friendly Computer! With Color and Music!", TI simply had a picture of the home computer on the face. Nothing about sprite graphics, more colors and voices than the Vic, and absolutely nothing about the 16 bit processor. TI was selling the computer to consumers the way you'd sell a calculator to an engineer: a small technical specs box.

Car Wars
I have just started an article about Car Wars, one of the video games mentioned in this article. The article is by necessity a stub, since I have based the information in it entirely on my childhood memories of playing the game. I could use some help from anyone who has more information. marbeh raglaim 05:56, 16 April 2007 (UTC)

calculator rematch
The TI/Commodore price war was Tramiel's revenge for TI pricing him out of the calculator market in the 70s. —Preceding unsigned comment added by 69.125.110.223 (talk) 15:54, 28 December 2007 (UTC)


 * Could you point us to a chart that illustrates the differences in prices between Commodore and TI calculators? 216.99.198.43 (talk) 20:01, 15 August 2009 (UTC)

Build quality
The article talks about the higher build quality of the TI - has this translated into a higher % of surviving examples?69.125.110.223 (talk) 18:16, 28 December 2007 (UTC)

Tomy Tutor redirect
Someone made a link about a port made on the "Tomy Tutor". When I clicked it, it brought me to the TI-99. I don't feel this is right because a redirect is like "The name of the article is not identical but it's the same thing". So, ok, the TT had TI-99 parts, but it wasn't THE TI-99. It's like redirecting a bicycle to a motorcycle because both have 2 wheels - but they are two very distinct objects. The TT really should have its own article and the redirect removed. -- Lyverbe (talk) 12:12, 30 May 2008 (UTC)

I agree - the "Tomy Tutor" was completely different in look to the TI-99. See http://oldcomputers.net/tomytutor.html vs. http://oldcomputers.net/ti994a.html. wiley14 (talk) 00:33, 12 August 2008 (UTC)

The Current Picture
Is a 99/4, not a 99/4a. Fix this, internet. --98.148.119.83 (talk) 07:32, 20 June 2009 (UTC)

ASCII on the TI/99 4A
I skimmed through the main article for information on the ASCII set used by the TI/99 4A. If the information is there, I missed it. Where is it? Maybe the article could be rearranged a little so the information is located at a major subheading? Could somebody knowledgeable please rewrite the main article to explain how Texas Instruments viewed the ASCII question? Did the computer have its own unique set of ASCII (for the sake of proprietarial control), and where were the character definitions? Were they in the GLOM, as I suspect? Was each letter effectively a sprite? How did the fonts differ from ATASCII and PETASCII, two standards that were fiercely competitive at the time? If the TI/99 4A was a 16 bit microprocessor, were the fonts laid out a letter at a time, 16 bits across by 8 or 16 bits deep, or did they follow some other standard? Could letters back then be a different color than the cursor?

Were there memory ports for controlling the shape, color, and location of the cursor? 216.99.198.43 (talk) 19:52, 15 August 2009 (UTC)


 * Characters are 8x8. You define a character in TI BASIC with a CALL CHAR command. CALL CHAR takes two parameters: the ASCII value of the character, and a string of 16 hexadecimal digits whose binary representation determines the pixel values. So CALL CHAR(65,"183C7EFFFF7E3C18") turns the letter A into a diamond shape. The cursor is a part of the character set itself, but is not in a character set affected by the CALL COLOR command (character code is either 30 or 31, I can't remember). Ntsimp (talk) 18:11, 4 September 2009 (UTC)


 * Thanks, that was very helpful.  216.99.198.97 (talk) 07:05, 9 September 2009 (UTC)


 * The 99-4/A supported the 256 character set, but AFAIK only 32 through 127 were standard ASCII. All 256 characters could be redefined. Each 8x8 pixel character could have two colors, but the colors had to be assigned to blocks of (IIRC) 16 characters. A problem cropped up with TI Extended BASIC because it reserved two blocks of characters for sprite functions. A TI BASIC program that redefined any character in those two blocks would not run in Extended BASIC. There were some programming workarounds for that to make such BASIC programs work in EX-BASIC and some 3rd party Extended BASIC's incorporated workarounds so that any BASIC program using the sprite reserved blocks would run. Unlike the regular character blocks for color assignments, every individual sprite could have its own color. IIRC the background color for sprites was always transparent. The 9918A did not have a true bitmap mode where every pixel was individually addressable for color. The "bitmap" mode was actually 32x192 1x8 pixel stripes. Each stripe could have two colors. Sprites were not affected by that limitation and could move around the screen without affecting (or being affected by) that limitation. Sprites could be 8x8 pixels or 16x16 pixels and a pixel doubled "magnified" version of each was available. That enabled a 4 step sprite "zoom" using only two sprite definitions. Bizzybody (talk) 08:03, 9 December 2010 (UTC)

Caps key on the TI 99/8
What does the "caps" key on the TI 99/8 do? It is the key on the upper left part of the keyboard, sort of located where the ESC key was, on similar computers of the era.

Is it similar to a caps lock key? Does it stay down when you press it, like a typewriter shift lock, or does it come up again, functioning more or less as a toggle? 216.99.198.126 (talk) 06:55, 18 August 2009 (UTC)

ROM replacements
How much addressing space was available on the ROMs? Were the ROMs socketed? If somebody opened up his TI-994A, did he have an easy way of unplugging the ROM and replacing it with an EPROM of his own design? 216.99.219.27 (talk) 02:23, 4 September 2009 (UTC)

The system ROMs had 8k of space from >0000 to >1FFF. They were soldered in place. For the most part they only contains the GPL interpreter. Most of the operating system was stored in GROMs. There's 6k of system GROM and 12k of TI BASIC GROM in the console. These chips often are socketed but GROMs were proprietary devices made by TI so not easily replacable. External GRAM Emulator devices would just override them by driving the bus harder than they could. —Preceding unsigned comment added by 76.203.123.41 (talk) 00:39, 24 November 2010 (UTC)

Release data
This article is not consistent: The photo says it was released in 1979. Directly underneath (and in the article) it says 1981. This needs to be made consistent. --Rehcsif (talk) 18:15, 3 April 2010 (UTC)


 * The article says that the TI-99/4A was released in 1981. The photo is of the predecessor of the TI-99/4A, the TI-99/4 (which is mentioned in the first paragraph of the article), which was released in 1979. HMishkoff (talk) 01:56, 4 April 2010 (UTC)
 * Sorry, completely missed that. As common as the 99/4A is, you'd think someone could post a picture of the actual machine, not the 99/4 which was much more rare.  --Rehcsif (talk) 16:36, 6 April 2010 (UTC)

Memory of the System
Someone keeps adding that that console has 16KB of CPU RAM /and/ 16KB of VDP RAM. This is not true for either the TI-99/4 or the TI-99/4A. They only have 16KB of VDP RAM and 256 bytes of CPU RAM. Please stop updating the article to say otherwise. You can clearly see this in any of the memory maps released from TI - check Editor/Assembler manual page 399 for one of many examples (explained on pages 398-399) - the only CPU RAM available besides scratchpad is the 32k memory expansion card. —Preceding unsigned comment added by 173.8.167.185 (talk) 23:17, 20 November 2010 (UTC)

Cleanup
This whole article needs a cleanup. Every section has had opinions inserted at random so many times that the whole thing reads somewhat random. The main sections should focus on the console as it was released and marketed, and all the custom user modifications, if they are worthy of mention at all, really should go in a separate section, to reduce confusion between what the machine originally was, and what people did with it. Perhaps we should check some of the other machine articles, are they as random? —Preceding unsigned comment added by 173.8.167.185 (talk) 23:31, 20 November 2010 (UTC)

An FPGA pin compatible 9918A replacement.
The F18A (for FPGA 9918A) project is nearing 100% original hardware compatibility. http://codehackcreate.com/archives/30 The prototype board is pin compatible with the 9918A (NTSC composite out), the 9928A (NTSC component out), 9929A (PAL component out) and TMS9919A (PAL component out, if such actually was made) Video Display Processor chips. How it does that is it does not use the pins that differ between the composite and component (color difference, works when connected to the common red, green and blue component video inputs) output versions of the chips. It has been tested in an MSX1, a TI-99/4A and a Colecovision without any changes to the F18A VHDL code. The F18A outputs video to a standard VGA 15pin connector, so there is no PAL or NTSC to bother with, its VGA. The creator of this plans to implement at least some features from the 9938 and 9958 VDP chips as well as some new features never available in that series of video chips. Aside from a perfect 9918A/9929A copy with VGA output + enhancements, another goal is to make the F18A much more affordable than any previous video upgrade for the TI-99/4A so that unlike previous video upgrades there will be many more TI owners able to use software that takes advantage of the enhanced features. The same should apply to use of the F18A in MSX1 computers. Bizzybody (talk) 08:22, 9 December 2010 (UTC)

Contemporary Use and Modern Hardware sections
Hi,

Today I have added some information about modern, contemporary use of the TI-99/4A, as it still enjoys a very active afterlife. I've also added a section briefly describing some of the modern hardware that is now available, or in development. It needs fleshing out more, and needs links/citations, but it's a start. Please have at it and go ahead and improve.

Regards

Mark Wills, 21 Jan 2011 —Preceding unsigned comment added by 178.97.147.31 (talk) 10:58, 21 January 2011 (UTC)

Reception...
Placed directly after Tech Specs seems to imply release reception, not some place-in-history review 25 years later. This IGN semi-factoid should be in legacy or something, not release reception, which discussed above anyway (as in sales, etc.)71.197.14.59 (talk) 11:06, 14 July 2011 (UTC)

99/4 release date
Someone changed the release date of the 99/4 from 1979 to 1977. I changed it back. I was working for TI in the Home Computer department until early 1979, and the product had not been released by the time I left, so I know that the 1977 date could not be possible. HMishkoff (talk) 00:46, 17 April 2012 (UTC)

Compatibility
The article doesn't seem to make this clear: software-wise, how compatible were the 99/4 and the 99/4A? --179.182.131.165 (talk) 17:38, 2 March 2019 (UTC)
 * They run almost all the same software. The only practical difference is the TMS9918A in the 4A. A program using the 'graphics 2' mode of that chip would run only on the 4A, but that was very rare. The notable example is Parsec. Ntsimp (talk) 20:54, 2 March 2019 (UTC)

Expand article to TI-99 in general?
While this article is currently titled TI-99/4A, it also is Wikipedia's only content on the original TI-99/4. I'm considering being bold in restructuring the article to be about the TI-99 family in general, which would cover both the 99/4 and 99/4A, but would also expand the sections at the bottom with what is known about the 99/2, 99/3, 99/4B(5), 99/7, 99/8, and other variants that were in development, prototypes built that still exist today, and in some cases advertised to the public before TI pulled the plug on the whole home computer division. At that point the article ought to be moved to "Texas Instruments TI-99." Is anyone opposed? –Mabeenot (talk) 16:40, 12 September 2020 (UTC)

Requested move 4 November 2021

 * The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review after discussing it on the closer's talk page. No further edits should be made to this discussion. 

The result of the move request was: Moved. (non-admin closure) Natg 19 (talk) 23:05, 11 November 2021 (UTC)

– The "TI" in "TI-99/4A" is an abbreviation of "Texas Instruments", so the current title seems redundant, and I'm pretty sure that no other company has made a notable "TI-99/4A". The target name for the main article already redirects here and always has since it was created in 2004. Similar articles, such as and other articles in Category:Texas Instruments calculators, are already at the shortened names. —&#8288;&#8202;&#8288;BarrelProof (talk) 01:31, 4 November 2021 (UTC) The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
 * Texas Instruments TI-99/4A → TI-99/4A
 * BBS software for the Texas Instruments TI-99/4A → BBS software for the TI-99/4A
 * Oppose, using the manufacturer name as well as the model seems to be the standard for home computers. VIC-20 redirects to Commodore VIC-20, PCjr redirects to IBM PCjr, etc... even though the shorter titles are unambiguous.  162 etc. (talk) 03:08, 4 November 2021 (UTC)
 * Those other article titles don't contain this company-name redundancy. What do you think of Texas Instruments 99/4A? —&#8288;&#8202;&#8288;BarrelProof (talk) 04:16, 4 November 2021 (UTC)
 * Please note that there is also a consistency argument in the other direction, per all the other TI-XYZ articles (again, see Category:Texas Instruments calculators, none of which use titles like –  and also HP-XYZ articles per Category:HP calculators, none of which use titles like, which doesn't even exist as a redirect). —&#8288;&#8202;&#8288;BarrelProof (talk) 14:13, 4 November 2021 (UTC)
 * Support – for some other computers, I'd also remove the manufacturer prefix, e.g. the VIC-20. For others, it's pretty much needed. Too much consistency is not a great thing. Dicklyon (talk) 04:25, 4 November 2021 (UTC)
 * Support: Right on cue, I knew someone was going to oppose this, citing consistency with inconsistency. Because WP:OTHERSTUFFEXISTS. Wikipedia is addicted to arbitrary corporate branding of every article title. They all need to be renamed. — Smuckola(talk) 05:27, 4 November 2021 (UTC)
 * I don't always mind including a company name in an article title about a product when it is unnecessary for disambiguation, but this one says TI twice in the name. That seems clearly 'beyond the pale'. —&#8288;&#8202;&#8288;BarrelProof (talk) 14:13, 4 November 2021 (UTC)
 * Yeah it is beyond the pale, and so is arbitrary corporate branding of titles. There are actual encyclopedic standards for names, WP:NATURAL and WP:COMMON, which are both violated by constantly prepending a corporate brand. What a person minds doing has no bearing anyway. That's not even how to disambiguate, but rather to add a parenthetical word, so all such things must be renamed. FYI! Thanks. — Smuckola(talk) 22:04, 4 November 2021 (UTC)
 * OK, but that's partly a discussion for another day (e.g., per WP:RMCI#Moves of other pages). At least for the moment, we are just noisily agreeing with each other. —&#8288;&#8202;&#8288;BarrelProof (talk) 22:59, 4 November 2021 (UTC)
 * Oppose - It's fine the way it is, the company name + model number. I've used to onw onw (my first computer) and I've never heard it called anything else. -  FlightTime  ( open channel ) 20:17, 4 November 2021 (UTC)
 * Your comment is illegible but "it's fine", and what you've heard, have no definition or bearing on Wikipedia. Even if you had somehow heard "Texas Instruments TI". — Smuckola(talk) 22:04, 4 November 2021 (UTC)
 * Boy, you're good Its just my opinion, thanks for sharing yours.  -  FlightTime  ( open channel ) 22:29, 4 November 2021 (UTC)
 * Support per nom.--Ortizesp (talk) 04:29, 5 November 2021 (UTC)
 * Support per nominator's reasoning. Tony (talk)  06:07, 6 November 2021 (UTC)
 * Support per nom and WP:CONCISE, and WP:PRECISE: we are only as precise as necessary. Over-disambiguation is almost never helpful. As for WP:CONSISTENT, the vast, vast majority of business/product names are not using redundant titles, so the actually consistent thing to do is make the few stragglers that are be consistent with the ocean of the rest of them, thus this move request to begin with.  — SMcCandlish ☏ ¢ 😼  18:13, 6 November 2021 (UTC)
 * Support based on successful precedents: Amiga instead of Commodore Amiga, ZX81 instead of Sinclair ZX81, Master System and Game Gear instead of prefixing them with "Sega". Dgpop (talk) 14:59, 7 November 2021 (UTC)

EDIT NOTICE misleading
I am confused by the edit notice for this article.

TI certainly did not knowingly release the source code, or any schematics besides the authorized copies in "PES and Console Technical Data".

Also, saying that any graphics, sounds etc are fair use, "do whatever", is not the precise meaning of "fair use".

Texas Instruments has never relinquished copyright on home computer materials, ROMs, etc (except for 3 limited exceptions in 1984). For instance, Plogue recently BOUGHT the rights from TI for the Parsec and other voices.

I am all in favor of attributing Texas Instruments on any material quoted! Credit where credit belongs.

However, the majority of source code/schematics that are out there were NEVER PUBLISHED. Therefore material from those can be removed from Wikipedia at any time. Exceptions would be published books that re-wrote this material, like TI Intern.

It's a conundrum that much of the best material about the TI-99/4A and cousins came from unpublished TI internal documents. The answers to many questions are recorded as oral history of TI employees, or unpublished documents--and ultimately, unpublished documents in the TI Archives at DeGolyer Library, Southern Methodist University. (which anyone can access. They are absolute treasures.)  Neither of these counts as "published" per Wikipedia policy.

The papers from the CB Wilson estate are an example of "floating around the Internet" and are the authoritative, primary source on the origin of the TI-99/4A. However, until they are archived (they ought to be preserved at DeGolyer with the rest of CB Wilson's papers) they can't even be published.

Texas Instruments has a process through the DeGolyer library for journalists to request permission to publish material from the archive.

FarmerPotato (talk) 20:05, 26 October 2022 (UTC)
 * Fair comment. I've requested that the notice be removed as it's unsourced, Google doesn't confirm it, and "has been released" is a very vague term in any case.  Chaheel Riens (talk) 05:47, 27 October 2022 (UTC)
 * have set that edit notice to expire this date 28 October 2022. If editors would like to see it or an updated version restored to the edit pages, then please join this discussion with your opinions and comments. [Link]  P.I. Ellsworth &thinsp;, ed.  put'r there 04:11, 28 October 2022 (UTC)

This edit notice has been deleted.  P.I. Ellsworth &thinsp;, ed.  put'r there 05:44, 8 November 2022 (UTC)