Alternate History: What Atari Could Have Done Differently?

[reifsnyderb]

4 minutes ago, cathrynm said:

C128 was hackish, but it did the job. To me, the logical progression from Atari 8-bit maybe would have been more like the Apple 2GS than the C128.  Compatibility with old 8-bit chips built into a single chip, but then a 16-bit 65816 mode with ST-ish/VGA-ish video modes and a SCSI port for hard drive, maybe a blitter. I do think Atari ST had basically a lot of the right decisions for features for the time.  Just GEM was a sloppy mess compared to MacOS, and Tramiel had burned a few too many bridges.

I believe there was some thought to go to the 65816 at some point.  The 1090XL had this card select line on it's edge connector that is always tied to ground at the side of the computer.  However, I think if a computer were to be connected that had the 65816 processor, this pin would have been used. 

 

I found an explanation in some of Atari's 1090XL documentation and have it added to my information, below:

 

Pin 13 Card Select

  Held low by the 1090 XL Expansion System. Atari planned this line to be an address select line for processors that could address more than 64k of memory.

  Connected to pin 37 of the PBI.

  No use of this line is known to exist.

 

I was stumped about the actual use of this line until I read about how addressing works on the 65816, when in 65816 mode.  This card select line now makes more sense.  I think this is some evidence that Atari engineers had plans to go to the 65816.

 

 

 

[zzip]

1 minute ago, reifsnyderb said:

I believe there was some thought to go to the 65816 at some point.

The "Super XEGS" the was under consideration would have been 65816 based  

 

 

[dely]

1. low-cost computer able to compete with the VIC20 (e.g. 600XL with 8 KB of memory) at the beginning of 1981 at the latest.
2. 800XL in early 1982 at the latest (to compete with the C64). 

I don't believe that the then quite big companies like Atari didn't have something like industrial espionage and didn't know what the competition was planning

3. Hiring some top programmers around 1982-1984 to develop AAA titles to stop the exodus of users to other computers. 
4. ANTIC+GTIA+POKEY in one chip around 1985/6 and a 16-bit machine based on the 65816 with 256 KB RAM, better graphics to display 80 columns, downward compatible with the 400/800/XL series instead of the XE which was totally unnecessary. 

[zzip]

3 minutes ago, dely said:

3. Hiring some top programmers around 1982-1984 to develop AAA titles to stop the exodus of users to other computers. 

There wasn't really an "AAA" scene back then, and no computer manufacturer did more than Atari to port hit games to their computers.   Atari was not just bringing the top arcade hits, but also funding Lucasfilm games to bring the next generation of videogames.    One problem was Atari was publishing all their popular titles to the competition via the AtariSoft label, and some of those ports were superior to the Atari hardware ports..   That didn't help!

 

But in that era of computing,  there was a move away from arcade style games and into brand new genres that only the power of a computer could unlock-- text adventures, simulations, 'contruction sets', RPGs, deeper sports sims.   A lot of these were disk-based and used the full memory of the machines.   Atari seemed committed to publishing on cartridge and ensure it worked even on the 16K models.   So third party was were the action was at rather than first party.

[phaeron]

A IIgs-style machine definitely would have extended the life of the platform, but IMO would still have been a dead end. Atari would have needed to evolve the platform for either productivity or game markets, both of which had problems.

 

Game market needed a machine with tiled playfields and sprites that were efficient to program. Not the funky display list + object line buffer stuff of the 7800. The NES could do higher resolution and more animated graphics with less RAM, and without needing a blitter or faster CPU. The next generation of consoles were adding more colors, more sprites, and more layers that were even harder to replicate compositing layers into a framebuffer.

 

Productivity needed better displays, better storage, more memory, but most importantly a better CPU. The 6502 was getting outdated and the 65816 was a dead end -- clunkier to program than both the 8088 and the 68000 and no viable route to 32-bit. The popular 65816 based machines ran at 2.8-3.5MHz; it was cool how far Apple stretched the platform with the IIgs and GS/OS, but I remember it being pretty slow. And eventually, even ~7MHz with cache was not going to cut it against the 68020 and 80386.

[Rybags]

Voice synth - I don't think it would have been much of a selling point.  Maybe some niche educational version but that market was probably sealed up by Apple by 1983 anyway.

 

The big tragedy was probably the XEM - on paper it looked like a great sound synth and could have boosted sales.

I agree, a IIgs type machine with more Ram, enhanced graphics+sound and extended OS (maybe GUI) would have been good.  But like the C65 they probably decided against it because it would obviously have impacted ST sales.

[cathrynm]

13 hours ago, phaeron said:

A IIgs-style machine definitely would have extended the life of the platform, but IMO would still have been a dead end. Atari would have needed to evolve the platform for either productivity or game markets, both of which had problems.

 

Game market needed a machine with tiled playfields and sprites that were efficient to program. Not the funky display list + object line buffer stuff of the 7800. The NES could do higher resolution and more animated graphics with less RAM, and without needing a blitter or faster CPU. The next generation of consoles were adding more colors, more sprites, and more layers that were even harder to replicate compositing layers into a framebuffer.

 

Productivity needed better displays, better storage, more memory, but most importantly a better CPU. The 6502 was getting outdated and the 65816 was a dead end -- clunkier to program than both the 8088 and the 68000 and no viable route to 32-bit. The popular 65816 based machines ran at 2.8-3.5MHz; it was cool how far Apple stretched the platform with the IIgs and GS/OS, but I remember it being pretty slow. And eventually, even ~7MHz with cache was not going to cut it against the 68020 and 80386.

For game systems, SNES was on 65816, Sega Genesis was on 68000, and I don't think it really made that much difference. Both systems had exactly the same games, sometimes, obviously built with the same art, probably from the same source code to both platforms.  SNES was a hint, that chip was available cheap in large quantities, but yeah, I don't think Antic/GTIA was on the path to evolving to take on something like Sega Genesis. 
On the PC side, pre-VGA, games were getting built from C rather than assembly, so anything with 16-color, 320x200, probably might have gotten a conversion if there were enough units out there.  It would have fit more in the Tandy 1000 category, I suppose.  An upgraded 8-bit with ST-level graphics would have obliterated ST completely, kind of my take. That basic the same games, but people can still plugin their old floppy drives.  Would have been a natural upgrade path to all those XEGS buyers, and old 8-bit people.  Atari would have hobbled together a GUI of some point, but GEM wasn't that great either. The problem is that it'd have taken slightly longer to design, and that delay might have been a deal killer.
On the applications side, Atari was just doomed because of Gem, is my take. The UI was sloppy, the rules for programming a compatible game that would survive upgrades were unclear. They had 68000 just like Mac but they never made the jump to the next CPU, not really.  When Atari went to 68030 too much stuff broke and those platforms all failed pretty hard. Tramiels never cared much about backwards compatibility, but by the late 80's and 90's, it mattered. I don't think they got that, and that's why they failed.

Edited April 2 by cathrynm

[zzip]

54 minutes ago, cathrynm said:

On the applications side, Atari was just doomed because of Gem, is my take. The UI was sloppy, the rules for programming a compatible game that would survive upgrades were unclear.

Atari got basically a 1.0 release of GEM,  yes it was rough around the edges, but so was Windows 1.0.    Problem was Atari kind of did the bare minimum to improve it over the years.   It didn't really get a desktop overhaul until 2.0 around 1990, five years after release.  And even then the changes were minimal compared to 3rd party desktops.   It didn't get multitasking until 1993.   The downside of having TOS in ROM is it isn't easy for most users to upgrade.

 

I think GEM could have evolved into something, but Atari was not the company to do it, like everything in the ST line, it fell behind what was happening in the PC world by 1990.

[cathrynm]

1 hour ago, zzip said:

Atari got basically a 1.0 release of GEM,  yes it was rough around the edges, but so was Windows 1.0.    Problem was Atari kind of did the bare minimum to improve it over the years.   It didn't really get a desktop overhaul until 2.0 around 1990, five years after release.  And even then the changes were minimal compared to 3rd party desktops.   It didn't get multitasking until 1993.   The downside of having TOS in ROM is it isn't easy for most users to upgrade.

 

I think GEM could have evolved into something, but Atari was not the company to do it, like everything in the ST line, it fell behind what was happening in the PC world by 1990.

I agree with all this. Old pre-Warner, maybe going into Warner-era Atari had some pretty sharp guys working over there, if they hadn't been idiots, they might have been a contender with Apple even, but by the time of the crash and the Tramiel era, all the software and UI brains had left the building. MacOS and the UI was the heart of why Apple became Apple and it's what mattered. I don't think the Tramiels had it in them to build this kind of software product.

[Pfaffa]

16 hours ago, phaeron said:

The popular 65816 based machines ran at 2.8-3.5MHz; it was cool how far Apple stretched the platform with the IIgs and GS/OS, but I remember it being pretty slow. And eventually, even ~7MHz with cache was not going to cut it against the 68020 and 80386.

Don't judge the 65816 speed by the Apple IIGS, the GS was intentionally run at a slower clock speed so that it wouldn't make the Mac look bad.

[pcrow]

43 minutes ago, Pfaffa said:

Don't judge the 65816 speed by the Apple IIGS, the GS was intentionally run at a slower clock speed so that it wouldn't make the Mac look bad.

So if Atari had gone with a 65816 system that could run 800/XL software, but had new graphics and sound chips, and a new GUI OS instead of the ST, they could have had a very successful system that could have had them in a strong position through the 80s.

 

But then they would have needed a follow-up 32-bit chip, say a 65832, or come up with something completely new to make it far in the 90s.  I'm not sure either was a path to success.

 

Still, that would have given us something really cool.  And it would have been nice if there were more to connect the 800 and the ST than the Atari logo.

 

Interesting ideas.  I wouldn't have thought of that as a real option without the above discussion.

[zzip]

5 minutes ago, pcrow said:

But then they would have needed a follow-up 32-bit chip, say a 65832, or come up with something completely new to make it far in the 90s.  I'm not sure either was a path to success.

It's not like the 32-bit TT/Falcon were massive hits,  and Atari exited the computer business shortly after Falcon, so lack of 32-bit in an 65816-based line would probably not have changed much.

 

The main problem was the 65816 didn't appear until late into the ST's design phase,  and the ST project started before Tramiel even knew he'd end up owning Atari, so there was no incentive to produce a backwards-compatible 16-bit computer anyway.

 

 

[phaeron]

5 hours ago, Pfaffa said:

Don't judge the 65816 speed by the Apple IIGS, the GS was intentionally run at a slower clock speed so that it wouldn't make the Mac look bad.

It doesn't matter why Apple chose 2.8MHz, if it was slow at that speed and the fastest popular 65816-based system wasn't clocked that much higher at 3.58MHz (SNES with fast ROM).

 

Part of the reason for this is that the 65816 requires faster memory than other contemporary 16-bit CPUs running at the same clock speed. The Amiga 500 ran a 7.14MHz 68000 on a memory bus with 280ns (3.58MHz) accesses and 150ns DRAM, with half the memory slots available for the chipset without slowing down the 68000. Effectively, a 7MHz 68000 doesn't need to access memory more often than the 1.79MHz 6502 since it only issues a read or write every 4 cycles. In the meantime, the BBC Micro needed 100ns DRAMs to run a 4MHz bus, which with video interleaving only ran the 6502 at 2MHz. I'm not sure that Apple could have released the IIgs at 8MHz at the start even if they wanted to, at least not without adding cache and making it more expensive.

 

Incidentally, out of curiosity I looked up the state of PC clones around the same time. The IIgs released at basically the same time as the Deskpro 386.

 

[zzip]

21 hours ago, cathrynm said:

I agree with all this. Old pre-Warner, maybe going into Warner-era Atari had some pretty sharp guys working over there, if they hadn't been idiots, they might have been a contender with Apple even, but by the time of the crash and the Tramiel era, all the software and UI brains had left the building. MacOS and the UI was the heart of why Apple became Apple and it's what mattered. I don't think the Tramiels had it in them to build this kind of software product.

Yeah,  my impression is the Tramiels were always trying to be on the leading edge of hardware (at a bargain price),  but never put the effort to be leading edge on software, they were always playing catch-up there, whether it be games, apps, OS, etc.    Unfortunately it turned out that software actually matters more than cool hardware specs

 

[Geister]

20 minutes ago, zzip said:

Yeah,  my impression is the Tramiels were always trying to be on the leading edge of hardware (at a bargain price),  but never put the effort to be leading edge on software, they were always playing catch-up there, whether it be games, apps, OS, etc.    Unfortunately it turned out that software actually matters more than cool hardware specs

 

If I recall correctly, Western Design Center had a 32 bit CPU waiting in the wings.  I believe they were waiting  for the 816 to get wider acceptance before bringing it out. https://downloads.reactivemicro.com/Electronics/CPU/WDC 65C832 Datasheet.pdf

Apple absolutely hobbled the 2GS to prevent competing with themselves.  At the time the 2GS already had a more finished GUI in color.  If it had more speed, the MAC would have died the early death it deserved.

The 8-bit 65C02 had a speed advantage over 16 bit 8088/6 CPUs, who's to say how that could have played out if WDC had customers for a next gen CPU?

Atari was at a disadvantage for advanced operating systems.  They had Gary Kildall's GEM system while Apple stole Xerox's WIMP interface.  Could Kildall have developed an interface better than the MAC's monochrome misappropriation?

But remember, all those who keep bringing up the ST and XE systems...in the alternate universe Tramiel stayed at Commodore and continued to build second-best systems (spec-wise, there's no accounting for cheap buyers), while Atari stood by its deals and paid for the Amiga development and distributed the Amiga as a 16/32 bit business system alongside it's 16/32 bit home Atari systems.

[zzip]

28 minutes ago, Geister said:

If I recall correctly, Western Design Center had a 32 bit CPU waiting in the wings.  I believe they were waiting  for the 816 to get wider acceptance before bringing it out. https://downloads.reactivemicro.com/Electronics/CPU/WDC 65C832 Datasheet.pdf

29 minutes ago, Geister said:

The 8-bit 65C02 had a speed advantage over 16 bit 8088/6 CPUs, who's to say how that could have played out if WDC had customers for a next gen CPU?

Yeah I'm sure that if the CPU was successful it would be extended.    But in the end it would still come up against the economies of scale in the PC clone market.   Yes the 8088 was very slow, and most certainly slower than the Atari 8-bit, but each generation of x86 CPU became more efficient and by the 386 & 486 it was pretty efficient.   The 68K line went all the way up to 68060, but that didn't help Atari.   By the time Atari released the 32-bit TT/Falcon, most people were buying PCs, and the PCs were giving more bang for the buck than Atari could deliver.

 

I think likely the same thing would have happened in a 65816 based computer line.   Eventually  the momentum of the PC market would be too much to compete against and resist.

 

37 minutes ago, Geister said:

Atari was at a disadvantage for advanced operating systems.  They had Gary Kildall's GEM system while Apple stole Xerox's WIMP interface.  Could Kildall have developed an interface better than the MAC's monochrome misappropriation?

My understanding is Atari got a fork of GEM, and Atari was responsible for updating it post ST launch without DR's involvement?    In the meantime PC GEM was trying to compete against Windows and not very well.  But at the same time, MSDOS apps dominated for another decade before Windows really took over the PC deskop.   So I don't know that you really needed to be Mac-like or even have a GUI system to have a competitive computer in the late 80s.   At the very least, Atari's GEM could have benefited from having a proper CLI,  the TTP (TOS takes parameters)  scheme was very awkward to use.

 

51 minutes ago, Geister said:

But remember, all those who keep bringing up the ST and XE systems...in the alternate universe Tramiel stayed at Commodore and continued to build second-best systems (spec-wise, there's no accounting for cheap buyers), while Atari stood by its deals and paid for the Amiga development and distributed the Amiga as a 16/32 bit business system alongside it's 16/32 bit home Atari systems.

Yeah in an alternate universe where Jack stays at Commodore who knows how it shakes out?  Atari would still have had to do a lot of cost cutting.   Maybe they take the TI route of deciding it's not worth competing against Jack and get out of the computer business.   Maybe they would have noticed the uptick in the console market and have been better prepared to keep Nintendo at bay.   Or maybe they jettison everything except for arcade because everything else is losing money,  or they find another buyer..

[+David_P]

On 4/2/2024 at 2:59 PM, cathrynm said:

I agree with all this. Old pre-Warner, maybe going into Warner-era Atari had some pretty sharp guys working over there, if they hadn't been idiots, they might have been a contender with Apple even, but by the time of the crash and the Tramiel era, all the software and UI brains had left the building. MacOS and the UI was the heart of why Apple became Apple and it's what mattered. I don't think the Tramiels had it in them to build this kind of software product.

Landon Dyer's blog (now only on archive.org) had a number of interesting stories about his time at Atari, including the day the Tramiels arrived.

 

https://web.archive.org/web/20200601213008/https://dadhacker-125488.ingress-alpha.easywp.com/

[Atari8guy]

The 80286 came out in 82, maybe a 16 bit system based on that, at that time, coupled with custom chips for graphics and sound and there is a way Atari stays relevant in the personal computer space.  

[dmsc]

Hi!

On 4/2/2024 at 11:02 PM, phaeron said:

Incidentally, out of curiosity I looked up the state of PC clones around the same time. The IIgs released at basically the same time as the Deskpro 386.

 

Yes, seems people tend to forget that the IIgs was very late, at that time Atari had already released the ST line. IMHO at that time the only sane CPU choice was the Motorola 68000, the 80286 a complicated mess and the "saner" 80386 was a lot more expensive.

 

Sadly, the 6502 was a dead end for productivity applications, I think that the only path forward would have been like the "apple-ii cards" for the Macintosh: a complete Atari 8-bit in a few integrated chips - a 65C02, a combination ANTIC-GTIA and perhaps an "extended" pokey.

 

About adding 80 column support, one way would be to add a font ROM (or static RAM) inside an extended GTIA and take the input from two ANTIC mode-15 lines to index the font data and produce the 80 columns - so you could simply replace the GTIA with this "80 column version" and a little software driver. Obviously you would loose the hardware scrolling and other features - but you could even support some "extended" graphics modes using the line buffer, for example 160x92 GTIA modes, or 320x92 with 4 colors.

 

Have Fun!

[phaeron]

I don't think the proposed W65C832 would have been effective either. It basically takes all of the weaknesses of the 65832 design and makes them worse.

 

What WDC basically did was just graft on an additional hidden E16 mode bit similar to the E/E8 bit, and add yet another set of modes. This means all of the same pains of switching between 8- and 16-bit widths, but with 32-bit width added on top. For instance, when dealing with both 8 and 16-bit quantities in 65816 native mode code, there are two main choices: use REP/SEP to switch modes at a cost of 3 cycles, or stay in 16-bit mode and emulate 8-bit accesses (e.g. LDA + AND #$00FF). But with the proposed 65C832, the E8 bit selects between 8/16 or 8/32 mode, so to switch between 16-bit and 32-bit memory/math, it's at least 4 cycles (CLC/SEC + XCE), and switching from 8-bit to 32-bit takes 7 cycles (REP + SEC + XCE). Compare that to the 68000 series, where all three widths are just encoded in the instruction, and the 80386 which just occasionally needs a size override prefix (66h).

 

On top of that, the 65C832 still only has an 8-bit data bus, so it's throttled like a 68008, except it's worse due to the lack of registers and memory access heavy design. LDA dp + CLC + ADC dp + STA dp to add two 32-bit variables takes 20 cycles while a 68008 can do ADD.L D4,D7 in 10 with a quarter of the bus cycle speed requirement.

 

Finally, none of the original problems of the 65xx series are addressed. If you need to access a strided 16-byte AoS structure, it takes the same old TYA / CLC / ADC #$10 / TAY sequence to do it, except now it's 11 cycles and 8 bytes. On other architectures you just do a single LEA or ADD.

 

[iguanaman3]

On 3/25/2024 at 10:49 AM, zzip said:

 

 

"Consoles are dead" - Because Tramiels believed this post-crash conventional wisdom (which they helped spread at Commodore),  they neglected the console market for a few years, giving Nintendo a huge opening to exploit and steal Atari's "king of gaming" crown.  Turns out that proprietary computers like the ST/Amiga were actually a dead end but proprietary consoles are still alive and well."

 

Yes! The attitude that computer games were a fad like slot car racing and would never again reach their previous levels was strong in the entire industry. It seems so ridiculous in hindsight. You have to wonder who was responsible for that myth and why did they do it.

 

 

 

Edited April 5 by iguanaman3
Addition

[xxl]

8-bit computers were dying and the development of this technology made no sense (the idea of producing accessories that increased the computer's capabilities - 1090 - was a signal that this line was coming to the end) extending the life of such equipment is also the cheapest way, i.e. releasing several killer apps. - just like it happens in the world of consoles.

And if there is a need to develop it, it happened - the Amiga designers predicted it... 680xx - but Atari overslept.

[Steril707]

Letting the Jay Miner and what became Amiga go was their big mistake.

Even more sad because the Amiga is simply the 16 bit Atari, while the ST is not.

 

[Geister]

11 hours ago, iguanaman3 said:

 

There were people predicting the death of Atari in 1984.  To most people Atari was the 2600.  They thought Atari was dead with the video game crash.

[reifsnyderb]

9 hours ago, xxl said:

8-bit computers were dying and the development of this technology made no sense (the idea of producing accessories that increased the computer's capabilities - 1090 - was a signal that this line was coming to the end) extending the life of such equipment is also the cheapest way, i.e. releasing several killer apps. - just like it happens in the world of consoles.

And if there is a need to develop it, it happened - the Amiga designers predicted it... 680xx - but Atari overslept.

Apple was still selling the Apple IIe Platinum into 1993.  Having an Atari with expansion slots would have been a relatively cheap way of extending the line.  I recently picked up an Apple II Plus.  It is amazingly primitive compared to the Atari.  The graphics are primitive.  The sound is nothing more than a CPU driven speaker.  The system has a 40 column display and doesn't even have a lower case character set.  Even the CPU still runs at 1 Mhz.  With the exception of the CPU, this can all be changed with expansion cards and a couple minor hardware hacks.  (i.e.  the shift key mod)  There was even a Z-80 card so the CPU wasn't really a limiting factor.

 

In my opinion, Atari screwed up by limiting the ability to expand the system. 

 

Some of the clues were there in the 800.  A 3rd party 80 column card was developed and sold.  Clearly, there was a demand for an 80 column display.  To me, anyhow, the XEP-80 was a half-baked solution for a serious problem.  The 1090XL 80 column card was much better...and never released.  The best solution would have been to have a higher end system, with an expansion bay, so as to have even more improved compatibility and upgrade capability than even the 1090XL would have had.  The 1450XLD had a big enough case for this.  The case design could have been used for expansion slots instead of floppy drives.

 

I understand that Atari also limited R&D....which resulted in the engineers that created Amiga to leave.  This was another major mistake.