The Texas Instruments 99/4: World’s First 16-Bit Computer | Hacker News Reader

The Texas Instruments 99/4: World’s First 16-Bit Computer

spectrum.ieee.org

54 pointsrbanffy43 comments

It was't the world's first 16bit computer. See:

https://en.wikipedia.org/wiki/16-bit

Maybe it was the first 16bit HOME computer... as mentioned in the text.

Hi -- I'm one of the editor's at IEEE Spectrum. You're absolutely correct: as the article does make clear, it was the first 16-bit HOME computer. We should have been more specific in the headline, our bad.

The Basic was slow for several reasons. 1. Double interpretation: As other already pointed out, the interpreter was written in an interpreted language, a kind of byte code if you prefer. 2. GROM routines: What added to the pain was that a lot of the routines necessary for the program were not in ROM but in TI specific GROM. All the "CALL something" (CALL CHAR, CALL HCHAR, CALL VCHAR, CALL SOUND, etc.) commands were GROM routines. GROM were a kind of serial ROMS that only TI manufactured (one of the reasons third parties had such a struggle with that machine). 3. VRAM: the worst offender why TI-BASIC (and TI-EXTENDED-BASIC) were so slow was that the basic console only had a wopping 256 bytes of memory. Yes, you read it right, the CPU only had 256 bytes of fast 16 bit memory. This memory was used by the system and were not accessible to the programmer. So where was the BASIC code then? It was stored in the video RAM and the CPU had no direct way to read and write to it. It always had to go through the TMS9918 video chip. This has 2 consequences, it was obviously slow and it clogged the video RAM. So unless one bought the 32K RAM expansion (which was outrageously expensive) or the mini-memory module (but then no Extended Basic was possible) it was impossible to use the graphic modes, making that incredibly advanced graphic chip a dud.

It's probably the strangest machine (aka f..d up) of the home computer era.

My parents got one of these when I was 7... It was a TI/994A, actually. I have fond memories of playing Parsec on it! I also taught myself Basic...

https://en.wikipedia.org/wiki/Parsec_(video_game)

And Parsec doesn't run on the original /4, due to its version of the TMS9918 graphics chip not having the bitmap mode.

i think TI basic was much easier than comodore basic; (still it really needed that extension cartridge)

https://en.wikipedia.org/wiki/TI_BASIC_(TI_99/4A)

Also it turns out that the interpreter was slow because it was written in an interpreted language, but it was good enough for simple games because they had built-in sprites.

Interesting that they have emulators/simulators for it.

http://www.99er.net/emul.shtml http://www.99er.net/win994a.shtml

No idea if this is true, but the meme was that you could only write BASIC for the TI99, assembly was forbidden somehow. For the most part, hackers avoided this system like the plague, despite all the advertisements starring Bill Cosby.

100% not true. There was an editor/assembler cartridge that let you program in ASM.

It had very little adoption and mindshare, though. All the assembly hackers were on the C-64, Apple II, and a bit later the PC.

Your other option was the Mini Memory cartridge, which added a scaled down version of Editor/Assembler as well as 8K of RAM and a battery backup.

It was an economical way to dabble in assembly without springing for the 32K RAM expansion and disk system.

I too started on TI BASIC, which taught me a lot, including ASCII, hex, binary, function calls, debugging the lists from magazines and how to type. I think I could type 'Call CLear' faster than anyone I knew! It's a pity I didn't get into ASM until years later when I got a spectrum.

I also had a TI99-4A ... It was a lot of fun to play with but I was also programming uProcessors in assembly language and that CPU's instruction set (and caveats) were so non-standard it was like learning it all over again.

I shudder to think how many hours I spent on that game. But even that paled in comparison to Tunnels of Doom.

http://ridingthecrest.com/edburns/classic-gaming/tunnels/

I wasted an alarmingly large percentage of a year or thereabouts on that game. And that was depite having the expansion chassis, floppy drives, the assembler card and Advanced Basic (all of which also ate a huge chunk of time).

I got good enough at Parsec to go infinite (earn lives faster than losing them) and roll over the score: http://dos486.com/misc/parsec.jpg

Fun fact: even though the manual says the sequence of level colors repeats after 16, it doesn't. It just stays yellow from level 16 onwards, through at least level 24 or so.

My first computer! I was 6. Started with TI invaders and ended up playing with assembly with minimemory a couple years later. Not unlike the Amiga, this machine was so in advance on its time for example with its hardware sprites.

This thing alone sparked my entire career.

If there are original TI engineers in the audience, it would be amazing to get a kind of AMA.

Hi Folks -- Stephen Cass here, one of the editors at IEEE Spectrum. Rather than replying to a bunch of comments individually, let me say here that the folks pointing out it wasn't the first ever 16-bit computer are absolutely correct. As the article quickly makes clear, it was the first 16-bit HOME computer: we should have been more precise in the headline, and that's our bad.

Speaking personally, I really loved having this article in Spectrum and reading everyone's comments, as I too was a 99/4A O.G. from the 1980s, (the machine was my first computer before I moved onto the BBC Micro). And word, yes, the system was squirrely, but I did at least learn to count in hex in my head trying to turn user-defined characters into poor-man's sprites in TI-BASIC

Here's a don't-miss article on the internal politics and strategy behind TI's early personal computing efforts:

https://books.google.ca/books?id=wy4EAAAAMBAJ&pg=PA63&lpg=PA...

The rest of that magazine issue is pretty interesting too, including the interview with Ann Winblad, fresh off her $15M exit. ("The male mystique was killed for me at age 2.")

Our school had one with pretty much everything in the way of expansions. It took up a lot of space on the table. I loved the voice synthesizer module which I fiddle around enough with to make it speak Dakota (Sioux) with a passable tone.

It may have been the first 16-bit home computer, but it certainly wasn't the first 16-bit computer. Or even the first 16-bit computer with a CPU on one chip.

And there were many, many bit-sliced 16-bit computers that predated it by years. Like the AYK-14: https://en.wikipedia.org/wiki/AN/AYK-14

https://en.wikipedia.org/wiki/Whirlwind_I

Whirlwind, circa 1950, was probably the 1st 16-bit computer.

Go GI!

Hi -- As I noted above, I'm one of the editor's at IEEE Spectrum. And you're absolutely correct: as the article does make clear, it was the first 16-bit HOME computer. We should have been more specific in the headline, our bad.

I love the TI-99/4A but what about the SuperPet 9000? It had a second 16-bit CPU you could use (the 6809) in _the same address space_ as the 8-bit 6502 CPU!

http://www.old-computers.com/museum/computer.asp?c=675

The 6809 was an 8 bit cpu (albeit an advanced one).

This was my first computer. I remember the huge RF modulator that came with it to hook up to my TV. I didn't get any games until quite a bit later (TI Invaders) I spent many hours trying to program tic-tac-toe from a book I got from the library.

The RF modulator is huge because it's also the power supply. Yet another odd quirk of this machine's design.

Oh wow, thanks for that! I've often thought of picking up one of these machines again... but I quickly realize that I'd really have little to do with it other than boot it up, and look at it. I think I'm probably more suited for an emulator.

The most enduring memory from TI-99 wasn't the games like MunchMan and Parsec. Instead, it was a specific oddity of the TI-BASIC language.

In the free TI-BASIC (built into ROM), the language had no boolean operators like "OR", "AND". Instead, what programmers did was use math tricks to simulate those booleans ... something like this:

Simulate "OR" boolean:

  IF (X=3) + (Y=5) THEN

Simulate "AND" boolean:

  IF (X=3) * (Y=5) THEN

As a child, it was extremely confusing to read those lines of code. Why in the world are we adding or multiplying nonsense like that?!? It turns out the programmers took advantage of any value greater than 0 as "TRUE", and 0 as false. Therefore, if either side evaluated to 0, multiplying the conditions would be FALSE.

However, you paid $100 extra for TI EXTENDED BASIC cartridge[1], that language had extra fancy features like the keywords "AND" and "OR". No more workarounds using "+" and "* " operators. That was progress in 1980s.

That early lesson in "inherent complexity" has affected how I view arguments today about "language simplicity". Not having things like "AND" "OR" makes the language syntax's formal specification simpler on a superficial level but it ends up making it more complex in user code.

[1] https://www.google.com/search?q=ti+extended+basic+cartridge&...

I've always heard that TI-BASIC conformed to a "strict ANSI BASIC" standard, whereas almost everyone else licensed Microsoft's version of BASIC.

Maybe that's why?

May be. TI BASIC was also limited in a lot of other ways. It didn't have ELSE and the THEN clause didn't allow for anything other than a line number as jump destination. Only one statement per line. So doing something as simple as 10 IF a=1 THEN v=10 ELSE v=50 would have to be written as 10 IF a=1 THEN 40 20 LET v=50 30 GOTO 50 40 LET v=10 50 ...

This was IMHO the worst restriction of that BASIC.

The only feature on the plus side of that BASIC was that it used the BCD floating format they also used in their calculator. So 13 digits precision with numeric range of 10^-99 to 10^99. When I switched later to the Apple II, I was shocked at the ridicule binary floating point format that couldn't even add 2 simple numbers correctly :-)

The TI-EXTENDED BASIC was an absolute must if you wanted to do serious programming. It enabled the sprites, had fancy stuff like named subroutines (that could even be called recursively) and a lot of restriction of the original BASIC were lifted.

My school had a 4A (which was pretty unusual in France back then) and it was the first time I touched a computer for programming.

2 friends and I were the only ones in school interested in it so we basically were in the "computer room" every lunch break :) We didn't have any recording device so it was just programming little stuff that will disappear at the end of the session. Had really great times !

I was 12. It was 33 years ago. Still developing.

My first computer was the 4A. I still have the ti-basic programming manual.

When I was 11 I used it to create fully-leveled NPC characters for dungeons and dragons.

I had one of these as a child in the '80s. My family had it hooked up to a black and white TV and I didn't know it could do color until seeing screenshots online in the early 2000s.

Something interesting that I learned recently is that the Manchester 'Baby' machine was a 32-bit computer, back in 1948.

There was a good article recently in El Reg about the TI community that still exists: https://www.theregister.co.uk/2017/02/27/community_builds_ar...

I wouldn't ordinarily pimp my own (poorly made) video, it's not even monetized, but I have the only TI-99/4 demonstration on YouTube if you're interested:

https://www.youtube.com/watch?v=oBgxOaUU7lQ

I'm sorry, but I genuinely don't understand what you mean. There are thousands of videos of people demonstrating use of this computer. Is there a specific property of a demonstration you are looking for?

The TI-99/4? Not that I've seen. 4A, yes. Maybe I'm in some kind of filter bubble, but I've searched far and wide for another demonstration of the 4.

Can't imagine there are thousands of videos of the 4 since there were barely thousands of them sold.

Didn't TI also had a dispute about the first microprocessor ? something internal they didn't really market or .. I forgot.

ps: I always loved the 99 design, especially with the side modules. Reminds me of old analog Meridian hifi modules

As an owner I can attest to the terribleness of the keyboard. I've used most of the bad chicklet keyboards out there - ZX, PCjr, Atari 400... this takes the cake.

I remember the Atari and TI being about the same level of horribleness. With the ZX and PCjr being unimaginably worse.

3 Mhz in 1981, very nice.

The games (Alpine, ...) I've looked up on YT look like C64 basic games though. Why was that?

Reply on news.ycombinator.com