Were the Atari ST's big for gaming or just the 8 bit line?

The Usotsuki · October 30, 2010

In the USA custom chips were a nice selling point given the C64 and A8 dominance in their respective period....and the ST had none so it didn't sell as well as here in the hobbyist market of the UK. Oh and PC in the home was acceptable for the yanks...but brits wouldn't piss on an 8086 machine in the 80s

ACT Apricot?

kool kitty89 · October 30, 2010

As far as failure of A8 in the UK goes it was mainly down to price/performance. The 800 was about £600 here which in 1982/3 was a lot of money. And the 400 did itself no favours with that keyboard and only 8k/16k

It seems the issues were far more than that, like programmer support, etc against the Speccy and C64, and not getting out low-cost hardware to better compete (which would indeed tie-in to the price issue obviously). From the pre XL perspective, the most straightforward route would be a UK/EU specific version of the standard BIOS/chipset but single board with no RF shielding, single (or no) cart slot, 2 joystick ports, built-in BASIC (and pack in detailed programming docs), and RAM socketed/soldered onboard with a single expansion port and probably a cheaper keyboard but one that was still acceptable and 16, 32, and 48k. (maybe a dome switch keyboard with decent key caps, though a full mechanical keyboard could have been OK) Beyond that they could have pushed the 1982 1200/600 designs for a more EU specific release as well, especially with the sleek 600 form factor. (especially if that had a full 64k and no RF shielding... though basically they needed the 800XL in both regions in '82, but even a direct cost-cut single board 400/800 with the same OS but simply adding BASIC and a consolidated board with no shielding would have helped big time)

The A8's main problem was simply down to competitive software support and marketing, without those it didn't have a chance at competing with the Speccy or C64.

The A8 was an older machine and thus at some disadvantages to the C64 (especially the lack of multiple indexed palettes for character cells -which the TMS9918 also did, but in VRAM, not dedicated CRAM -and speccy had the attribute cells for the framebuffer -not the same as char cells though), but if properly utilized it could still have held its own and did have advantages as well. (much larger color palette, true bitmap modes for software rendering -potential for great versions of Elite and Starglider, faster CPU, 4 4-bit linear DACs to work with for software PCM and 3 interval timers to facilitate that using interrupts -very well suited to the fast interrupts of the 650x arch, albeit you do have more overhead for being dependent on palette reloading to get more colors on BG and sprites -it's a bit odd that there aren't many game/homebrew demos for sample/MOD player stuff on the A8 given all the ST/Speccy stuff and even some C64 examples, and for the pure sound generation you did have a good bit of flexibility for square or variable pulse waves and some possibility of triangle/saw stuff, though for really good results you had to pair for 16-bit freq range instead of 8, it's universally better than the SN76489 at least and better with some trade-offs compared to the AY8910)

But I'm sure we don't need another A8 vs C64 discussion on the technical side of things. (let alone another general conflict for those liking the A8+amiga and not C64/ST or A8+ST but not Amiga/C64 or ST+C64+Amiga but not A8 )

The biggest thing would simply have been getting more established from release up through 1983 when home computers really took off. (the A8 line should have been an attractive alternative to the BBC Micro and with lower end models more contemporary to the VIC -with the very low end with the ZX80/81 of course)

The success of the ST in the early days in the UK comes down to it being a nice traditional home computer, sure it wasn't that great on custom chips side BUT 2/3 biggest home computer sellers had no custom chips anyway...only the C64 had it all and at a reasonable price.

Well pretty much all had some form of custom chips (even the Speccy had the video ASIC) aside from those which truly became clone/defacto standards with any custom hardware becoming off the shelf or simply custom hardware that was later opened to the mass market. (ie the TMS9918 was custom back when TI first used it in-house, but it later became an off the shelf part used by 3rd parties, and then the clone side of things with the IBM PC line etc)

But the ST was great in the sense that it was good enough in a number of ways, a great bang for the buck, and at a reasonable price (Amiga was a good bang for the buck, but with more bang and more buck so to speak pushing it into a higher price bracket -in part due to the higher end form factor in general, but I'd imagine even an A500 alike back in '85/86 would have been a good bit more expensive than the ST) So it wouldn't be surprising if a lot of Speccy users moved on to the ST in the late 80s. (common/popular acceptable performance and most affordable in its class by far -had the Loki project panned out, things might have been different, though honestly that project would have been even more interesting to get pulled into Amstrad and developed into a console in leu of the GX-4000 -the Loki team did go on to develop Flare 1, then the Slipstream ASIC based on that, then the Jaguar )

Granted there's a number of trade-offs to in-house and off the shelf chipsets: for a custom chipset there's no overhead from buying a pre-existing product as you own the IP (and if you have your own chip fab facilities, that cuts that out as well -but even outsourcing like AInc/Corp did was a big help over off the shelf offerings), you also have possibilities for things that not only would be very expensive off the shelf, but many times not available at all. Only when you really get massive production of off the shelf stuff does it really get competitive and make the R&D overhead of in-house product less efficient. (the Odyssey 2 went all fo the shelf with Intel parts, but I'll bet it was at a significant disadvantage to cost/performance compared to the possibilities of an in-house chipset -5200 vs Colecovision should be a similar case but I think there's more mitigating factors there)

Thing is in the UK we liked computers for lots of reasons, programming was a big thing as was doing no games type tasks.

Yes, which is another major issue that came up in the context of the A8 being marketed as a closed box appliance computer, not catering to programming as such. (plus the EU/UK home programming/hobby market was a bit different than the US one, especially with the really low-end side of things and apparently greater interest from younger users and going beyond hobby and into actual commercial games in some cases at young ages) Sinclair set a standard by packing in Documentation, Commodore learned quickly, but Atari lagged without eden a built-in BASIC interpreter until the XL line in '83/84.

There's been an ongoing (rather long-winded) discussion related to this on Sega-16: http://sega-16.com/forum/showthread.php?t=13468&page=12 (including a parallel argument of if the Speccy could have been marketable in the US at a lower end bracket than the C64 and taking advantage of EU software support)

But what I found particularly interesting was the apparently weak software development tool support from Atari in '80-83. (especially compared to the pack-in docs of competition)

And as for games, I wouldn't bother to load up 75% of all ST OR Amiga games today because once cowboys like Ocean and US Gold treated it as a business to fleece us with crap ports of expensive arcade licences rushed out the door it was all over. Stuff like Marble Madness for Amiga and Pawn/Gauntlet 1 on ST were revolutionary at the time. Very few later games came close to showing the same sort of attention to detail/quality programming for me. Shadow of the Beast/Lotus being a little blip in a sea of shit for the Amiga technically.

That's what happens to all popular platforms, tons of crappy cache-in stuff in general: plenty of VCS, NES, Genesis, SNES, PSX, PS2, PC, Wii, etc games like that.

In the USA custom chips were a nice selling point given the C64 and A8 dominance in their respective period....and the ST had none so it didn't sell as well as here in the hobbyist market of the UK. Oh and PC in the home was acceptable for the yanks...but brits wouldn't piss on an 8086 machine in the 80s

I doubt the custom chips had much to do with it, but more so the general competition and lesser price sensitivity of the US market. Neither the ST or the Amiga did amazingly well and the Amiga mainly held on longer because it remained technically competitive longer (ie until VGA+SB+386 PCs became common in the early 90s), both got overshadowed by PCs though and the ST more quickly with its lesser advantages (a late 80s 286+Adlib+EGA machine would be fairly competitive), and even with contemporaries you had the Tandy 1000 line making up the biggest direct competition in lower/mid-range computers with PC compatibility and graphics/sound more competitive with the ST and far ahead of common CGA based clones. (especially with no sound cards coming on the scene until '87 -and you had EGA becoming lower end at that same time though TGA graphics were generally equal to EGA games anyway as they tended to stick to the default palette 320x200 modes -in fact TGA may have been advantageous as it uses packed pixels rather than planar graphics -which would also give an advantage over software rendering on the ST, but you had weaker X86 CPUs to counter that for lower end 8088/8086 versions -still 7.16 MHz vs only the 4.77 of many clones, and with the 8 MHz 8086 you got a little boost, and much much more with the 8 or 10 MHz 286 models -of course it later went up from there as well to 25 MHz 386SX machines and pack-in VGA cards plus an 8-bit DMA sound channel with sampling up to 48 kHz -and Cyrix had the 486SLC clip on upgrades for 386SXs as well, and it of course had standard ISA expansion slots) And you did get the deskmate GUI as well, and all out a year earlier than the ST or Amiga in the US. (it's sort of an indicator of what the PC Jr might have been had IBM made it a standard PC simply aimed at a lower-end market with game oriented graphics/sound -ie standard ISA and external interface ports) Not sure if any had options/provisions for FPUs though.

The lower-end contemporary early ('85/86) T-1000s were far cheaper than contemporary Amigas and fairly close to the ST's price I believe, but with less features (usually 256 kB RAM, 7.16 MHz 8088, and one 360 kB 5.25" floppy) and usually coming with a color composite monitor (and like CGA all TGA resolutions had colorburst disabled modes allowing RGB quality monochome/grayscale graphics on a composite monitor -particularly necessary for any text based stuff with the 80 column mode being totally useless in color composite -as is the 640x200 graphics mode unless you want artifact colors, with 320x200 being boarderline for text and OK for games -I don't think artifacting was taken advantage off for the TGA modes, but if used like CGA, that would have allowed roughly 256 pseudo colors at an effective 160x200), and you got PC compatibility and expandability. (CPU upgrades would be tougher for the 8088 and to some extent 8086 machines though -you could get a bit of a boost with some of NECs 88/86 compatbiles)

Hell, even back with the original release in 1984 the Tandy 1000 was advertised at $1,199 and the bundles usually included a monitor. (though contrary to most depictions in adds, it would almost certainly be advertising the price with 1 FDD, not 2)

I think you also mused before on the lack of a higher-end more professional looking ST early on (ie box+keyboard form factor -even without a faster CPU or expandability), something that didn't come until the MEGA 2 years later (and still w/out CPU upgrade -which would have been particularly nice due to the delay of the BLiTTER, but could have been done from the start with a higher-end version -expansion provisions would have been nice too). It's a bit of a paradox: the Amiga lacked a lower-end console form factor until '87 with the more professional box only, while the ST was the opposite. (though it's a bit odd that neither the MEGA nor the A2000 upgraded the CPU to a 12/16 MHz one -and maybe an optional FPU) Granted, Atari did fake the professional look by putting a pair of disk drives under the monitor in leu of a box.

I think you also pointed out that even with the lower-end models, faster CPUs would have been more attractive than the blitter (or AGA) as all (non timing sensitive) games would benefit from it. (but they did need to have more comprehensive upgrades too, especially to compete with VGA... in fact I think you could argue that a 320x200x256 color mode with 12-bit RGB would have been a good deal more important than the blitter... especially in combination with a faster CPU and especially if it used chunky bitmap graphics like VGA -not sure if the TT did, I know AGA didn't short of the CD32- more higher-res modes would be useful too, like 640x200x16 color, but 320x200x8bpp would have been most significant both in meeting VGA/MCGA's 13h and consequently better competing with the Amiga -for general graphics and especially art/paint programs as well as games -even without a blitter chunky 8bpp rendering should have been much faster or at least easier to manage than 4 bitplanes and it would have greatly facilitated VGA ports as well -especially given both would be pure software rendered stuff -and somewhat akin to the ST blitter, PC games tended to not take advantage of the 2D acceleration offered by many SVGA cards of the early 90s with the primary use being for windows acceleration, but sound expansion was another story as those caught on fast with many big-name developers though Adlib/SB became the most widely used... a shame the ST cart slot didn't have an audio-in line actually it's more of a shame that it hadn't been a general expansion port -in that case, the blitter could have been more practical in general as well and the video used the main system bus, so it wouldn't mean adding a much more pins for that either, and in that case the blitter would be an easier upgrade than an new VDC/SHIFTER as you'd need a separate video jack or analog RGB input lines on the expansion port with a toggle to bypass onboard video -not genlock)

That's really it I think, the ST wasn't expandable enough for the US, the very same issue Atari Inc learned after trying to push the A8 as an "appliance computer" (which got worse with the 1200 and fixed with the 600/800XL's PBI and with the 1090XL module at least), had the ST had a general expansion port in the same vein (with external expansion module/rack) it quite possibly could have been a different story. (especially with a complementary higher-end system with internal slots of the expansion system already onboard). Of course they'd definitely want to keep the low-end machine too (especially for Europe), but a general purpose expansion port in place of the unused cart port shouldn't have hindered that either. (as it was, the cart port was used more as an expansion port than anything else...

Hell, even the VIC-20 and C64 had (fairly) general purpose expansion ports. (sometimes referred to as cartridge ports though... albeit many game consoles had a lot of added signals to the cart ports for expansion though not always taken advantage of -the Genesis even had genlock capabilities with the YS line)

Though the PET was really an appliance computer in the same sense if not more so and the MAC was/is as well. (at least the all-in one MACs, not the MAC II or modern MAC Pros or such) Sort of ironic given that the Apple II had such an emphasis on expandability with that as a major advantage over the A8. (even if the A8 generally still had more features than many expanded Apple machines and was much cheaper -it couldn't do 560 pixel across though, so max was 40 column text vs 80 column expansions of the Apple II -albeit rather useless with a composite monitor -even 40 column in the A8 can be rather iffy in composite/RF -Y/C monitors would obviously solve that, or had they offered modes with colorburst disabled like CGA -though for 800/1200 users you could simply plug the luma cable into a composite video input or even have a switch to go between color and clean grayscale, though if you wanted to use a TV you'd need to mod internally to the RF modulator input or use an external RF modulator)

PAL doesn't have nearly as bad a problem with color artifacting at those resolutions given the higher colorburst signal. (newer composite encoders and comb filters help a lot too)

Edited October 30, 2010 by kool kitty89

oky2000 · October 30, 2010

In the USA custom chips were a nice selling point given the C64 and A8 dominance in their respective period....and the ST had none so it didn't sell as well as here in the hobbyist market of the UK. Oh and PC in the home was acceptable for the yanks...but brits wouldn't piss on an 8086 machine in the 80s

ACT Apricot?

A few clueless yuppies had them but not the millions of people from C64/A8/Amstrad/Spectrum backgrounds. Just check out all the EGA/CGA crap before about 1991 and you see why in the 80s PC gaming was a dead end and usually inferior graphically to the ST (and sonically too a lot of the time with PC Speaker or ADLIB options only) let alone the Amiga, which by the time of the A500 at 500 then 400 pounds was the slam dunk. Unless you needed to take work home to do stuff on Lotus 123 or Dbase or WordPerfect it was a simple choice.

emkay · October 30, 2010

I'd bet that I wrote the 1st inter Atari game( 800 and ST connected with a serial device) , but I dislike the ST in much senses.

The A800 had so much phantastic advantages, but the jameels ( ) decided to make a cheap machine.... wahhhh.... what could have been even cheaper than to use the hardware developed in the own house? ANTIC, GTIA, and POKEY were available. Just make them 16 bit was given. But that lazy ex-commies didn't care a fart and used this YM - and this unworthy Shifter Chip.

The ST series had this "slight" advantage of the straight 8MHz clocked 68K CPU, while the AMIGA had to reduce the clocking to the Chipset's needs(7.19MHz) . At the End, you couldn't even dare to compare both machines, if you're clear at mind, because the AMIGA is similar to TWO STs with some minus and some plus .....

Ofcourse, compared to the 8 bit Atari games, the ST was way better in resolution and in games that needed a faster CPU. But, there often wa a sour feeling involved.

Edited October 30, 2010 by emkay

kool kitty89 · October 30, 2010

In the USA custom chips were a nice selling point given the C64 and A8 dominance in their respective period....and the ST had none so it didn't sell as well as here in the hobbyist market of the UK. Oh and PC in the home was acceptable for the yanks...but brits wouldn't piss on an 8086 machine in the 80s

ACT Apricot?

A few clueless yuppies had them but not the millions of people from C64/A8/Amstrad/Spectrum backgrounds. Just check out all the EGA/CGA crap before about 1991 and you see why in the 80s PC gaming was a dead end and usually inferior graphically to the ST (and sonically too a lot of the time with PC Speaker or ADLIB options only) let alone the Amiga, which by the time of the A500 at 500 then 400 pounds was the slam dunk. Unless you needed to take work home to do stuff on Lotus 123 or Dbase or WordPerfect it was a simple choice.

That's a big difference with Europe and the UK... PC gaming pretty much followed up the wave of mass popularity of the C64 (ie '86/87 onward).

Again, you DID have the Tandy 1000 from 1984 onward and that was almost certainly the best bang for your buck as far as PC compatibles prior to 286 EGA machines becoming affordable in the late 80s. (and even so it wasn't until VGA that that outpaced the 1984 Tandy graphics, but EGA tended to get a lot more support once it took off -and on top of the Tandy's SN96489 -not quite as good as the ST's YM2149- and the DAC added to late 80/90s models, but you had ISA cards to go with -had the DAC been supported better, you'd only really need an Adlib card as the joyports and DAC were there, but the SB's DAC was supported far more often, at least by the early 90s -without expansion you more or less had a mono equivalent to the STe's sound, or single channel mono rather as the STe had 2 DMA channels with panning) Again, I'm not positive, but I think the TGA graphics were all chunky pixel like CGA, so lacking the overhead of planar graphics. (not sure about the odd interleaved scanline arrangement of CGA though)

And Adlib sound is superior in every respect to the ST's onboard sound though it might not have always been used well (compare crappy ST music to crappy adlib music though...). Albeit the YM2149 did give a good route for CPU driven PCM playback while the YM3812 didn't (had they added a simple resistor DAC it would have addressed that though, but that's going to eat up CPU resource just like software playback on the ST and the SB added actual DMA audio). The Covox Sound Master also had DMA audio (mapped to also be compatible with parallel port DAC sound) and an AY-3-8930 (significantly enhanced successor to the AY8910 with envelopes expanded to all 3 channels, variable pulse wave, more flexible noise generation, etc) and added 2 DE-9 atari style digital joyports (using the 8-bit I/O ports of the chip), however that card was unpopular with the Soundblaster being released the same year with DMA sound and the superior FM synthesizer (compatible with Adlib) and the IBM analog joyport (granted, inferior to the Atari style ones). Had there been earlier cards using the AY8910 to thus be backwards compatible with, that would likely have changed things. (actually rather odd that that didn't happen in 84-86 or even the simpler SN76489 -and in either case a simple 8-bit resistor latter would have been useful and way better than PWM for sample playback, though so would using the PSG itself -either would be less intensive and way better sounding and the PC already had interval timers to allow interrupt driven PCM).

But from the perspective of backwards compatible upgrades to the AY8910, Yamaha did one better with the YM2203 in 1985 with full PSG (SSG) and I/O functionality and a 40 pin DIP, but adding 3 4-op FM channels (like the YM2151 or YM2612 vs the simpler 2-op of the 9 channels in the YM3812) with the only catch being the need for an external DAC (a tiny 8-pin DIP) but so does the YM3812. (the super low-end YM2413 didn't though, and neither did the YM2612 -which also supported an 8-bit direct write mode for the DAC and integrated interval timers) So that would have been a good option for the ST as well, probably something service centers could even offer as an upgrade (let alone had the cart port had audio in lines -even a separate audio out could have been put on the expansion cart), and they were already getting their sound chips supplied by Yamaha.

I'd bet that I wrote the 1st inter Atari game( 800 and ST connected with a serial device) , but I dislike the ST in much senses.

The A800 had so much phantastic advantages, but the jameels ( ) decided to make a cheap machine.... wahhhh.... what could have been even cheaper than to use the hardware developed in the own house? ANTIC, GTIA, and POKEY were available. Just make them 16 bit was given. But that lazy ex-commies didn't care a fart and used this YM - and this unworthy Shifter Chip.

The ST series had this "slight" advantage of the straight 8MHz clocked 68K CPU, while the AMIGA had to reduce the clocking to the Chipset's needs(7.19MHz) . At the End, you couldn't even dare to compare both machines, if you're clear at mind, because the AMIGA is similar to TWO STs with some minus and some plus .....

...

They did initially plan to put AMY in the ST, but that was going too slow and they had to drop it. (I wonder how cost effective it would have been to stick a 6502+POKEY onboard with some interface logic to facilitate that... especially if POKEY's keyboard I/O could have been used in leu of the Hitachi scanner and maybe POKEY's SIO as well and maybe even the interval timer functionality, but you'd still need 8 I/O lines for the parallel port; even without trying to make a backwards compatible machine that might have made for a nice coprocessor set-up, especially for sampled audio using the 6502 and POKEY's DACs)

It probably would have been a lot more cost effective to simply use a YM2203 in place of the YM2149, at least if Yamaha had them in suitable volume at the time. (the PC8801 had it in '85) It also added 2 programmable interval timers, so that might have offset cost too. (eliminating the need for some other timers in the ST)

Edited October 30, 2010 by kool kitty89

oky2000 · October 31, 2010

As far as failure of A8 in the UK goes it was mainly down to price/performance. The 800 was about £600 here which in 1982/3 was a lot of money. And the 400 did itself no favours with that keyboard and only 8k/16k

It seems the issues were far more than that, like programmer support, etc against the Speccy and C64, and not getting out low-cost hardware to better compete (which would indeed tie-in to the price issue obviously). From the pre XL perspective, the most straightforward route would be a UK/EU specific version of the standard BIOS/chipset but single board with no RF shielding, single (or no) cart slot, 2 joystick ports, built-in BASIC (and pack in detailed programming docs), and RAM socketed/soldered onboard with a single expansion port and probably a cheaper keyboard but one that was still acceptable and 16, 32, and 48k. (maybe a dome switch keyboard with decent key caps, though a full mechanical keyboard could have been OK) Beyond that they could have pushed the 1982 1200/600 designs for a more EU specific release as well, especially with the sleek 600 form factor. (especially if that had a full 64k and no RF shielding... though basically they needed the 800XL in both regions in '82, but even a direct cost-cut single board 400/800 with the same OS but simply adding BASIC and a consolidated board with no shielding would have helped big time)

The A8's main problem was simply down to competitive software support and marketing, without those it didn't have a chance at competing with the Speccy or C64.

The A8 was an older machine and thus at some disadvantages to the C64 (especially the lack of multiple indexed palettes for character cells -which the TMS9918 also did, but in VRAM, not dedicated CRAM -and speccy had the attribute cells for the framebuffer -not the same as char cells though), but if properly utilized it could still have held its own and did have advantages as well. (much larger color palette, true bitmap modes for software rendering -potential for great versions of Elite and Starglider, faster CPU, 4 4-bit linear DACs to work with for software PCM and 3 interval timers to facilitate that using interrupts -very well suited to the fast interrupts of the 650x arch, albeit you do have more overhead for being dependent on palette reloading to get more colors on BG and sprites -it's a bit odd that there aren't many game/homebrew demos for sample/MOD player stuff on the A8 given all the ST/Speccy stuff and even some C64 examples, and for the pure sound generation you did have a good bit of flexibility for square or variable pulse waves and some possibility of triangle/saw stuff, though for really good results you had to pair for 16-bit freq range instead of 8, it's universally better than the SN76489 at least and better with some trade-offs compared to the AY8910)

But I'm sure we don't need another A8 vs C64 discussion on the technical side of things. (let alone another general conflict for those liking the A8+amiga and not C64/ST or A8+ST but not Amiga/C64 or ST+C64+Amiga but not A8 )

The biggest thing would simply have been getting more established from release up through 1983 when home computers really took off. (the A8 line should have been an attractive alternative to the BBC Micro and with lower end models more contemporary to the VIC -with the very low end with the ZX80/81 of course)

I was simplifying things for sure, it was generally an issue of cost in the early VIC20/Spectrum/C64 days around 1982/83. There were probably 40 different computers for people to choose from in the UK at this time and most people looked at cost of machine and cost of games/availability of games. Atari A8 games were either on cart or disk and imported. This stacked up badly against 5.99 cassette games for Sinclair and 6.99 games for C64.

ZX Spectrum, very basic but just good enough with colour and beeper sound. £125 for 16k

C= 64 which was light years ahead in almost all aspects but cost a whopping £325/299

Acorn BBC Micro was somewhere between Spectrum and C64 but cost £600 for 32k I think.

Spectrum was cheap and chearful with 100s of games coming out weekly in 1983. C64 had serious graphical/sonic advantages and a real keyboard and 64kb RAM (a seriously big amount on launch day) and many games weekly by 1983. Acorn BBC had not many games, cost a fortune too but was the one your kids used at school so toffee nosed parents bought it for 'educational' reasons for their children.

There wasn't much room for anyone else really, and to be honest Commodore were very lucky they had VIC-II and SID designed by engineers payed a regular salary AND they produced them at cost price. Atari could never undercut Commodore on custom hardware production costs. 64kb RAM was expensive to build into any machine in 1982 and was a massive risk.

Also there were lots of tricks people started to discover even within a couple of years, the C64 char mode was a god send for colourful games using more than the 4 colours, and sprite multiplexing and then sprites in the border all added a certain glitz that none of the other 8bits had and this is pretty much why it just sold like hot cakes for a decade. The A8 was quite difficult to code and the secrets of the various quirks to exploit were much slower in arriving, by the time they did A8 was old news and ST was the Atari talk of the day in the mid 80s.

The success of the ST in the early days in the UK comes down to it being a nice traditional home computer, sure it wasn't that great on custom chips side BUT 2/3 biggest home computer sellers had no custom chips anyway...only the C64 had it all and at a reasonable price.

Well pretty much all had some form of custom chips (even the Speccy had the video ASIC) aside from those which truly became clone/defacto standards with any custom hardware becoming off the shelf or simply custom hardware that was later opened to the mass market. (ie the TMS9918 was custom back when TI first used it in-house, but it later became an off the shelf part used by 3rd parties, and then the clone side of things with the IBM PC line etc)

But the ST was great in the sense that it was good enough in a number of ways, a great bang for the buck, and at a reasonable price (Amiga was a good bang for the buck, but with more bang and more buck so to speak pushing it into a higher price bracket -in part due to the higher end form factor in general, but I'd imagine even an A500 alike back in '85/86 would have been a good bit more expensive than the ST) So it wouldn't be surprising if a lot of Speccy users moved on to the ST in the late 80s. (common/popular acceptable performance and most affordable in its class by far -had the Loki project panned out, things might have been different, though honestly that project would have been even more interesting to get pulled into Amstrad and developed into a console in leu of the GX-4000 -the Loki team did go on to develop Flare 1, then the Slipstream ASIC based on that, then the Jaguar )

Granted there's a number of trade-offs to in-house and off the shelf chipsets: for a custom chipset there's no overhead from buying a pre-existing product as you own the IP (and if you have your own chip fab facilities, that cuts that out as well -but even outsourcing like AInc/Corp did was a big help over off the shelf offerings), you also have possibilities for things that not only would be very expensive off the shelf, but many times not available at all. Only when you really get massive production of off the shelf stuff does it really get competitive and make the R&D overhead of in-house product less efficient. (the Odyssey 2 went all fo the shelf with Intel parts, but I'll bet it was at a significant disadvantage to cost/performance compared to the possibilities of an in-house chipset -5200 vs Colecovision should be a similar case but I think there's more mitigating factors there)

Ahh when I talk about custom chips I mean not using the CPU to scroll the screen, move graphics memory, create moving graphics like Sprites or Blitter Objects on C64 and Amiga respectively. As I said Commodore had MOS make VIC-II/SID and then bought Amiga computers and again made the chipset themselves. Although they paid a lot more to get the Amiga chipset than it cost them to end up with the C64.

From this point of view the ST didn't have anything bespoke/unique to it. There was no scrolling/animation in hardware and the sound chip was more or less identical to that from a lowly Spectrum 128k or Amstrad CPC 8bit computer as far as consumers were concerned.

Thing is in the UK we liked computers for lots of reasons, programming was a big thing as was doing no games type tasks.

Yes, which is another major issue that came up in the context of the A8 being marketed as a closed box appliance computer, not catering to programming as such. (plus the EU/UK home programming/hobby market was a bit different than the US one, especially with the really low-end side of things and apparently greater interest from younger users and going beyond hobby and into actual commercial games in some cases at young ages) Sinclair set a standard by packing in Documentation, Commodore learned quickly, but Atari lagged without eden a built-in BASIC interpreter until the XL line in '83/84.

There's been an ongoing (rather long-winded) discussion related to this on Sega-16: http://sega-16.com/forum/showthread.php?t=13468&page=12 (including a parallel argument of if the Speccy could have been marketable in the US at a lower end bracket than the C64 and taking advantage of EU software support)

But what I found particularly interesting was the apparently weak software development tool support from Atari in '80-83. (especially compared to the pack-in docs of competition)

The Timex computers were not significantly cheaper than the Commodore 64 in the USA? Sure software would be a problem and require importing (so reversing the situation the A8 was hindered by in the UK in the very early 80s). Personally Americans had far more disposable income, the Spectrum was unique to the cash starved Brits who WANTED cheap and cheerful so doubt it would make a difference, the Spectrum just has too many compromises to be worth the few quid you save.....unless you don't have a few quid extra and the alternative is nothing at all to play with or wait 2 years and buy a second hand computer etc

And as for games, I wouldn't bother to load up 75% of all ST OR Amiga games today because once cowboys like Ocean and US Gold treated it as a business to fleece us with crap ports of expensive arcade licences rushed out the door it was all over. Stuff like Marble Madness for Amiga and Pawn/Gauntlet 1 on ST were revolutionary at the time. Very few later games came close to showing the same sort of attention to detail/quality programming for me. Shadow of the Beast/Lotus being a little blip in a sea of shit for the Amiga technically.

That's what happens to all popular platforms, tons of crappy cache-in stuff in general: plenty of VCS, NES, Genesis, SNES, PSX, PS2, PC, Wii, etc games like that.

It's rare to see a Megadrive game which doesn't have 64 colours, isn't using full screen parallax and does not use all the features of the soundchip with FM AND AM sound routines for music and speech/samples etc. Compare Gauntlet 4 from Megadrive to Gauntlet 2 on Amiga. The Amiga game looks same as the ST version, the ST version is INFERIOR to Gauntlet 1 in all but horizontal scrolling (screen size smaller, tiles are 2 colour, sprites smaller, animation slower, less speech/samples etc). If you look at Gauntlet 1 on ST and compare to Gauntlet 4 on Megadrive you get an accurate gauge of the difference in power of both. Now run Gauntlet 2 Amiga side by side with Gauntlet 4 and a n00b on the street would think the Amiga is graphically inferior to an ST running Gauntlet 1!

Sadly very few games on ST/Amiga are anything other than the bare minimum, and total rush jobs by cheap ass UK Software houses spending the cash on arcade licenses and then hiring rubbish talentless development groups to hash it all up. On SNES/Megadrive/PC-Engine AKA Turbo Grafix the norm is very polished code and lovely graphics and sound from talented people. Gotta love those industrious Japanese games coders!

In the USA custom chips were a nice selling point given the C64 and A8 dominance in their respective period....and the ST had none so it didn't sell as well as here in the hobbyist market of the UK. Oh and PC in the home was acceptable for the yanks...but brits wouldn't piss on an 8086 machine in the 80s

I doubt the custom chips had much to do with it, but more so the general competition and lesser price sensitivity of the US market. Neither the ST or the Amiga did amazingly well and the Amiga mainly held on longer because it remained technically competitive longer (ie until VGA+SB+386 PCs became common in the early 90s), both got overshadowed by PCs though and the ST more quickly with its lesser advantages (a late 80s 286+Adlib+EGA machine would be fairly competitive), and even with contemporaries you had the Tandy 1000 line making up the biggest direct competition in lower/mid-range computers with PC compatibility and graphics/sound more competitive with the ST and far ahead of common CGA based clones. (especially with no sound cards coming on the scene until '87 -and you had EGA becoming lower end at that same time though TGA graphics were generally equal to EGA games anyway as they tended to stick to the default palette 320x200 modes -in fact TGA may have been advantageous as it uses packed pixels rather than planar graphics -which would also give an advantage over software rendering on the ST, but you had weaker X86 CPUs to counter that for lower end 8088/8086 versions -still 7.16 MHz vs only the 4.77 of many clones, and with the 8 MHz 8086 you got a little boost, and much much more with the 8 or 10 MHz 286 models -of course it later went up from there as well to 25 MHz 386SX machines and pack-in VGA cards plus an 8-bit DMA sound channel with sampling up to 48 kHz -and Cyrix had the 486SLC clip on upgrades for 386SXs as well, and it of course had standard ISA expansion slots) And you did get the deskmate GUI as well, and all out a year earlier than the ST or Amiga in the US. (it's sort of an indicator of what the PC Jr might have been had IBM made it a standard PC simply aimed at a lower-end market with game oriented graphics/sound -ie standard ISA and external interface ports) Not sure if any had options/provisions for FPUs though.

Even in early 93/late 92 a 486SX 25mhz with no sound and the slowest ISA SVGA card cost £1000+ from the most dodgy clone makers. In the UK around sometime between Amiga 500 and 1200 Amstrad tried to launch the Sinclair PC10 which was an all-in-one style 8086 CGA machine with 512kb/360kb drive and blurry TV modulator to compete at around 400-500 quid. It failed, the worst computer format launch in the history of computers. Nobody wanted CGA in the time of C64 or EGA in the time of Amiga. Shadow of the Beast was impossible on PC technically before 1990 and even then in 1992 a machine to play Shadow of the Beast would cost £1500+ Ditto Super Stardust in 1995...A1200 £299 Pentium 120mhz £1000+ Ditto Lotus II Turbo with A600 costing £299 486DX2-66 costing + Roland sound card AND Soundblaster = £1500+

The point is Commodore never ever did anything useful with the Amiga chipset and refused to licence it to Amstrad either. The A500 was as close as they came to building a spiritual successor to the C64 but technology was moving at an ever faster pace. CGA to EGA was longer than EGA to MCGA/VGA and it all snowballed. Ditto 8088/86/V30 to 80286 affordable PCs to the later years of breathtaking pace of ever increasing Mhz of 486/Pentium on a monthly basis! Commodore were trying to compete with 25mhz CPUs and Atari with 16mhz CPUs in Falcon with 16bit RAM not 32bit! People gave up anyway by the time Commodore went tits up. People got fed up with a slow CPU. If you can't make an awesome custom chip to run DOOM on a 14mhz computer then give up on custom chip development and ramp up the CPU speed to match the brute processing power of competing machines (ie in 1993 that would be 386DX 25/33 competing with 14mhz crippled 68020 with no 32bit fast ram included off the shelf...madness you get 200% CPU speed with just 512kb of non chip ram on an A1200...the machine is effectively running at the speed of an 8mhz 020 as sold by Commodore!)

The lower-end contemporary early ('85/86) T-1000s were far cheaper than contemporary Amigas and fairly close to the ST's price I believe, but with less features (usually 256 kB RAM, 7.16 MHz 8088, and one 360 kB 5.25" floppy) and usually coming with a color composite monitor (and like CGA all TGA resolutions had colorburst disabled modes allowing RGB quality monochome/grayscale graphics on a composite monitor -particularly necessary for any text based stuff with the 80 column mode being totally useless in color composite -as is the 640x200 graphics mode unless you want artifact colors, with 320x200 being boarderline for text and OK for games -I don't think artifacting was taken advantage off for the TGA modes, but if used like CGA, that would have allowed roughly 256 pseudo colors at an effective 160x200), and you got PC compatibility and expandability. (CPU upgrades would be tougher for the 8088 and to some extent 8086 machines though -you could get a bit of a boost with some of NECs 88/86 compatbiles)

Hell, even back with the original release in 1984 the Tandy 1000 was advertised at $1,199 and the bundles usually included a monitor. (though contrary to most depictions in adds, it would almost certainly be advertising the price with 1 FDD, not 2)

In the UK we never had such machines really, and most crappy 8086 'affordable' PCs had 640k RAM, 5.25" disks from C64/A8 days and CGA graphics in crap looking cases. Amstrad probably made the best of the worst. Their answer to the ST/Amiga was the Sinclair PC10...see this monstrosity here

And really that was it, somehow the UK buyer who accepted really compromised rubbish like the Spectrum was having none of this typing at a DOS prompt, EGA coloured scratchy soundblaster based PC crap until Atari/Commodore dropped the ball with Falcon/AGA Amiga respectively. And even then they had to go bust first too before games companies really pushed PC gaming. I know in the USA this was not the case and PC gaming was normal by the early 90s. To be honest though, Wing Commander 1 is a shit game Doom was the first game I would have bought a PC for.

I think you also mused before on the lack of a higher-end more professional looking ST early on (ie box+keyboard form factor -even without a faster CPU or expandability), something that didn't come until the MEGA 2 years later (and still w/out CPU upgrade -which would have been particularly nice due to the delay of the BLiTTER, but could have been done from the start with a higher-end version -expansion provisions would have been nice too). It's a bit of a paradox: the Amiga lacked a lower-end console form factor until '87 with the more professional box only, while the ST was the opposite. (though it's a bit odd that neither the MEGA nor the A2000 upgraded the CPU to a 12/16 MHz one -and maybe an optional FPU) Granted, Atari did fake the professional look by putting a pair of disk drives under the monitor in leu of a box.

I think you also pointed out that even with the lower-end models, faster CPUs would have been more attractive than the blitter (or AGA) as all (non timing sensitive) games would benefit from it. (but they did need to have more comprehensive upgrades too, especially to compete with VGA... in fact I think you could argue that a 320x200x256 color mode with 12-bit RGB would have been a good deal more important than the blitter... especially in combination with a faster CPU and especially if it used chunky bitmap graphics like VGA -not sure if the TT did, I know AGA didn't short of the CD32- more higher-res modes would be useful too, like 640x200x16 color, but 320x200x8bpp would have been most significant both in meeting VGA/MCGA's 13h and consequently better competing with the Amiga -for general graphics and especially art/paint programs as well as games -even without a blitter chunky 8bpp rendering should have been much faster or at least easier to manage than 4 bitplanes and it would have greatly facilitated VGA ports as well -especially given both would be pure software rendered stuff -and somewhat akin to the ST blitter, PC games tended to not take advantage of the 2D acceleration offered by many SVGA cards of the early 90s with the primary use being for windows acceleration, but sound expansion was another story as those caught on fast with many big-name developers though Adlib/SB became the most widely used... a shame the ST cart slot didn't have an audio-in line actually it's more of a shame that it hadn't been a general expansion port -in that case, the blitter could have been more practical in general as well and the video used the main system bus, so it wouldn't mean adding a much more pins for that either, and in that case the blitter would be an easier upgrade than an new VDC/SHIFTER as you'd need a separate video jack or analog RGB input lines on the expansion port with a toggle to bypass onboard video -not genlock)

I have always confessed that had Atari opted to replace the original ST with an 8mhz/16mhz switchable CPU ST rather than the 8mhz STE with inferior [to the Amiga] updates that nobody wanted to rewrite 1000s of lines of code for things would have been different up until around 1990/91. 91 was when there were VGA versions of arcade/action PC games not just EGA. I don't mean simulation stuff from USA I mean conversions of games like Lotus III etc. 16mhz 68000 Instead of wasting all that money on the STE enhancements that were never really used in exclusive games and isolated the ST/STM/STF/SFTM owners AND split the market for the software houses who had to decide what game to do and which format and then which bloody model of ST/STE etc. Just play games on STeem emulator at 16mhz and you can have a lot of fun. Gauntlet 1 scrolls better and is awesome and Lotus II is approaching Amiga levels of smoothness.

Obviously had Commodore or Atari sold as well as their first big hits (C64 and VCS respectively) then ploughed the money back into a next gen update sooner then things would not have gone the way of 14mhz A1200, 16mhz Falcon or 66mhz PC as your only choices!

That's really it I think, the ST wasn't expandable enough for the US, the very same issue Atari Inc learned after trying to push the A8 as an "appliance computer" (which got worse with the 1200 and fixed with the 600/800XL's PBI and with the 1090XL module at least), had the ST had a general expansion port in the same vein (with external expansion module/rack) it quite possibly could have been a different story. (especially with a complementary higher-end system with internal slots of the expansion system already onboard). Of course they'd definitely want to keep the low-end machine too (especially for Europe), but a general purpose expansion port in place of the unused cart port shouldn't have hindered that either. (as it was, the cart port was used more as an expansion port than anything else...
Hell, even the VIC-20 and C64 had (fairly) general purpose expansion ports. (sometimes referred to as cartridge ports though... albeit many game consoles had a lot of added signals to the cart ports for expansion though not always taken advantage of -the Genesis even had genlock capabilities with the YS line)

Though the PET was really an appliance computer in the same sense if not more so and the MAC was/is as well. (at least the all-in one MACs, not the MAC II or modern MAC Pros or such) Sort of ironic given that the Apple II had such an emphasis on expandability with that as a major advantage over the A8. (even if the A8 generally still had more features than many expanded Apple machines and was much cheaper -it couldn't do 560 pixel across though, so max was 40 column text vs 80 column expansions of the Apple II -albeit rather useless with a composite monitor -even 40 column in the A8 can be rather iffy in composite/RF -Y/C monitors would obviously solve that, or had they offered modes with colorburst disabled like CGA -though for 800/1200 users you could simply plug the luma cable into a composite video input or even have a switch to go between color and clean grayscale, though if you wanted to use a TV you'd need to mod internally to the RF modulator input or use an external RF modulator)

PAL doesn't have nearly as bad a problem with color artifacting at those resolutions given the higher colorburst signal. (newer composite encoders and comb filters help a lot too)

All the ST had to do in 1985 was kill the hype coming from Apple with the original Macintosh, a terrible terrible machine. I think the styling Atari chose for the ST was a mistake in 1985 in some ways. It wasn't cheap enough at £799 inc Mono monitor to be bought by people who would be alienated by a nice sexy 3 box design like the original Megas and it was looking too much like an 8 bit computer their children played games on than a serious tool for someone who would buy a £799 machine with all the performance of a £2500 Macintosh and appreciate GEM, DMA hard disks, excellent serious software potential.

They probably should have launched the 520STM in parallel with the 520ST! It's not like it is any different to the original ST and modulators cost peanuts so that would have allowed the early adopter to get right in there in 1985 with a shiny new computer at something like £150 less, home computer owners didn't want to pay £999 for 520ST+SC1224 and the mono monitor is useless to someone trading up from an old 8bit colour computer. SM124+ST was only really going to work for Macintosh comparisons. To be honest the ideal perfect launch of the ST would be the 520STM to home computer elite users (like the ones who bought a BBC Micro for £800 with disk drive in 1982) and a 512kb machine in the style of the Mega ST to break into serious/business user market. Sell the 520STM @ £550 in 1985 with SF354 bundle (or actually launch some bloody software on cartridge!) and then sell the 'Mega 0.5 ST' with SM124 and SF354 for £799ish to people who liked the Mac but weren't stupid enough to pay £2500 for a 128k monochrome piece of crap.

I have always suspected that the Atari ST that Shiraz Shivji designed was based on plans for a 16bit successor to the PET before he left Commodore. If you look at the ST in this light you will see it fits really well. It has the right mix of colour and resolution for the time to help with doing business graphics as a successor to the PET would need. Needs to be easy and fuss free to use because business users are not loyal and don't appreciate difficult to use systems so GEM would have been perfect. Make sure you can swap data with PC disk format check. Make the machine faster than comparable computers of the time with a clean fast 8mhz 68000 design with 4mb addressable RAM. Include excellent hard disk interface on board with DMA port. Sprinkle some unique features like 3 channel sound and 512 colour palette too. You see how it makes sense? Had the ST been sold by a company called anything but Atari the business world would have chewed your arm off to get hold of one. The 70hz 32khz non interlace 640x400 high res coupled with the DMA port for hard disks and PC DOS compatible disk format would have worked very well for a PET replacement from Commodore. Hell Commodore in the 70s was very highly respected by business customers to be honest. How things change lol

PS I should point out, in case anyone doesn't know, the original 520ST or 520STM for me is the most stylish all in one computer ever made design wise, and my comments are merely the perception this form factor gives to potential business users of your machine who all have used computer terminals hooked to mainframes, Apple Mac/Lisa or IBM compatibles and who only see this type of all in one design on 'kids computers' they purchased with 8bit technology for their kids to play space invader clones on a few years ago

svenski · October 31, 2010

That's a big difference with Europe and the UK... PC gaming pretty much followed up the wave of mass popularity of the C64 (ie '86/87 onward).

PC Gaming didn't take off until the early nineties. The mass wave of popularity of the C64 you talk about moved on to the Amiga and ST, the 16-bit generation. not PCs.

oky2000 · October 31, 2010

I'd bet that I wrote the 1st inter Atari game( 800 and ST connected with a serial device) , but I dislike the ST in much senses.

The A800 had so much phantastic advantages, but the jameels ( ) decided to make a cheap machine.... wahhhh.... what could have been even cheaper than to use the hardware developed in the own house? ANTIC, GTIA, and POKEY were available. Just make them 16 bit was given. But that lazy ex-commies didn't care a fart and used this YM - and this unworthy Shifter Chip.

The ST series had this "slight" advantage of the straight 8MHz clocked 68K CPU, while the AMIGA had to reduce the clocking to the Chipset's needs(7.19MHz) . At the End, you couldn't even dare to compare both machines, if you're clear at mind, because the AMIGA is similar to TWO STs with some minus and some plus .....

Ofcourse, compared to the 8 bit Atari games, the ST was way better in resolution and in games that needed a faster CPU. But, there often wa a sour feeling involved.

I don't think you understand business very well. Jack bought Atari from Warner with his PERSONAL FINANCES. Had he not done that ATARI WOULD BE DEAD IN 1984 simple as that.

Commodore only made business errors AFTER Jack left. He made the PET reality, he made the C64 happen with his purchase of MOS with his profits and he had no more money or MOS technologies engineers to make custom chips and spend 2 years to make Amiga-II. Everything after that EXCEPT Amiga 1000 is flawed from Commodore. Commodore 16/Plus4 = crap. Commodore 128/128D is overpriced underpowered. Amiga 2000 same shit as A1000 2.5 years later. Commodore LCD machine would have cleaned up but Commodore never released it because some idiot listened to Tandy who told him it would never sell, it was finished and complete!

A8 was dead, Making ANTIC/GTIA/POKEY cost Atari Warner 2-300% MORE than MOS making SID/VIC-II for Jack @ Commodore. Dead end technology, and by now the C64 coders were making awesome music on SID and getting ALL 16 colours on screen with 32 sprites multiplexed in 160x200 in 1985/86.

Also Jack rescuing Atari from bankruptcy/shut down by Warner cost him most of his own personal money, did you expect them to design 16 bit custom chips for $10 in 6 months? Every day he owned Atari he was paying off debts racked up thanks to the rubbish sales of 5200 and A8 worldwide after Warner ran it into the ground, he made it to 1995 from nothing, second time in his business life.

A new computer was needed FAST, GEM is a genius OS superior to Mac, the ST was technically far superior to the Mac but cost 65% LESS. He did a great job in 1985 with the money he had left over rescuing Atari from the creditors. STE was a waste of time though too little too soon. Alienated the public AND the games developers and split his market and increased production costs making STE AND STFM together in factory.

AMY was nice but probably cost more than Amiga Paula chip. Antic/GTIA were old news, no place for mono sprites and 160x200 4 colour graphics on 16bit £799 computers in 1985/86

oky2000 · October 31, 2010

That's a big difference with Europe and the UK... PC gaming pretty much followed up the wave of mass popularity of the C64 (ie '86/87 onward).

PC Gaming didn't take off until the early nineties. The mass wave of popularity of the C64 you talk about moved on to the Amiga and ST, the 16-bit generation. not PCs.

In Europe yes, but in the USA it is exactly as Kool Kitty says. Have a scan through US magazines like Computer Gaming World and you will see it was C64/Apple II/A8 and then it was C64 and PC and then mostly PC EGA gaming in the USA. Amiga and ST were small time blips in the world of America. After C64 people went to Nintendo NES or PC gaming really. In contrast flicking through Computer and Video Games from 1984 to 1988 you will see it went from Spectrum/C64 to Amiga/ST. NES didn't sell here because we didn't want C64 graphics and £40 cartridge games. EU Nintendo 8bit sales = fail It wasn't until the SEGA Megadrive in 1990 onwards we were even interested in consoles again since the time of Coleco/VCS in the UK that's for sure. Nintendo's first hit here was the puke-o-rama screened Gameboy+Tetris in 1991.

PC VGA 386 and above gaming started to take off around 1992 because Commodore dropped the ball with the slowest ever 256 colour screen mode EVER made and a CPU crippled to 50% speed thanks to ALL Chip RAM on the Amiga 1200 Then there was CD32 which couldn't do SNES/Megadrive quality games but cost £399! And then they went bust which helped a bit too for PC gaming.

Atari made mistakes too, the Falcon was crippled with slow memory, then Jaguar came out at the wrong time and Playstation/Saturn/3DO killed it with their texture mapped 30FPS 3D games. Oh and the 520STFM RAM price problem around 1989/90 when the ST had to go up from £299 to £399 ....just when Commodore decided to make the Amiga 500 £100 cheaper to the magic £399...same price as the ST

oky2000 · October 31, 2010

In the USA custom chips were a nice selling point given the C64 and A8 dominance in their respective period....and the ST had none so it didn't sell as well as here in the hobbyist market of the UK. Oh and PC in the home was acceptable for the yanks...but brits wouldn't piss on an 8086 machine in the 80s

ACT Apricot?

A few clueless yuppies had them but not the millions of people from C64/A8/Amstrad/Spectrum backgrounds. Just check out all the EGA/CGA crap before about 1991 and you see why in the 80s PC gaming was a dead end and usually inferior graphically to the ST (and sonically too a lot of the time with PC Speaker or ADLIB options only) let alone the Amiga, which by the time of the A500 at 500 then 400 pounds was the slam dunk. Unless you needed to take work home to do stuff on Lotus 123 or Dbase or WordPerfect it was a simple choice.

That's a big difference with Europe and the UK... PC gaming pretty much followed up the wave of mass popularity of the C64 (ie '86/87 onward).

Again, you DID have the Tandy 1000 from 1984 onward and that was almost certainly the best bang for your buck as far as PC compatibles prior to 286 EGA machines becoming affordable in the late 80s. (and even so it wasn't until VGA that that outpaced the 1984 Tandy graphics, but EGA tended to get a lot more support once it took off -and on top of the Tandy's SN96489 -not quite as good as the ST's YM2149- and the DAC added to late 80/90s models, but you had ISA cards to go with -had the DAC been supported better, you'd only really need an Adlib card as the joyports and DAC were there, but the SB's DAC was supported far more often, at least by the early 90s -without expansion you more or less had a mono equivalent to the STe's sound, or single channel mono rather as the STe had 2 DMA channels with panning) Again, I'm not positive, but I think the TGA graphics were all chunky pixel like CGA, so lacking the overhead of planar graphics. (not sure about the odd interleaved scanline arrangement of CGA though)

And Adlib sound is superior in every respect to the ST's onboard sound though it might not have always been used well (compare crappy ST music to crappy adlib music though...). Albeit the YM2149 did give a good route for CPU driven PCM playback while the YM3812 didn't (had they added a simple resistor DAC it would have addressed that though, but that's going to eat up CPU resource just like software playback on the ST and the SB added actual DMA audio). The Covox Sound Master also had DMA audio (mapped to also be compatible with parallel port DAC sound) and an AY-3-8930 (significantly enhanced successor to the AY8910 with envelopes expanded to all 3 channels, variable pulse wave, more flexible noise generation, etc) and added 2 DE-9 atari style digital joyports (using the 8-bit I/O ports of the chip), however that card was unpopular with the Soundblaster being released the same year with DMA sound and the superior FM synthesizer (compatible with Adlib) and the IBM analog joyport (granted, inferior to the Atari style ones). Had there been earlier cards using the AY8910 to thus be backwards compatible with, that would likely have changed things. (actually rather odd that that didn't happen in 84-86 or even the simpler SN76489 -and in either case a simple 8-bit resistor latter would have been useful and way better than PWM for sample playback, though so would using the PSG itself -either would be less intensive and way better sounding and the PC already had interval timers to allow interrupt driven PCM).

But from the perspective of backwards compatible upgrades to the AY8910, Yamaha did one better with the YM2203 in 1985 with full PSG (SSG) and I/O functionality and a 40 pin DIP, but adding 3 4-op FM channels (like the YM2151 or YM2612 vs the simpler 2-op of the 9 channels in the YM3812) with the only catch being the need for an external DAC (a tiny 8-pin DIP) but so does the YM3812. (the super low-end YM2413 didn't though, and neither did the YM2612 -which also supported an 8-bit direct write mode for the DAC and integrated interval timers) So that would have been a good option for the ST as well, probably something service centers could even offer as an upgrade (let alone had the cart port had audio in lines -even a separate audio out could have been put on the expansion cart), and they were already getting their sound chips supplied by Yamaha.

Thing is here in the UK around 1988/87 the next big thing was the Acorn Archimedes with it's RISC based 32bit CPU design and 256 colour graphics and 8/16 channel stereo 8bit sound (8 channels with panning control so same a 16 Amiga channels)

This thing was lightning fast, the speed of a 25mhz 020 when Commodore launched the 7mhz 68000 A500 and Atari the 520STE with 8mhz 68000. All the UK magazines raved about this machine, and it was a seriously powerful machine. But sadly it died the same fate as the A8 did in the UK. Expensive advanced machine with very little support from software houses.

Like I said the PC was nowhere here in the UK, PC technology was just too expensive (A PC soundcard from Roland cost as much as an Amiga 500 new!) and there was just no interest before 1990/91 because games were EGA. Rocket Ranger, Defender of the Crown, It Came from the Desert were ALL inferior to the Amiga version and even graphically inferior to the ST version. The screenshots did PC gaming no favours, and Cinemaware was where it was at in the late 80s end of story.

On sound specifically, most early ST games had sampled sound, Guantlet 1...Technocop....Dungeon Master etc. So whilst the Yamaha chip in the ST was worse than ADLIB in some ways....if you had ADLIB your sound effects were all coming from PC speaker...barely an improvement on a 1982 ZX Spectrum's sound. ADLIB is FM sound I believe, like the SEGA Megadrive, it's not really something you can compare. What I do know is, the ST was more flexible but as time went on and software houses got more greedy less and less sampled sounds were in games and things got worse. You can't play Amiga MODs on a PC without a mid 90s wavetable card like GRAVIS or AWE32 etc, you can on an ST though...bit scratchy sure but you can do it Trouble was in the 80s you need ADLIB for music and Soundblaster for effects. Expensive...2 sound cards for 1 machine. And both were inferior to Amiga and pretty much the C64 SID too. There was talk of getting SID onto an ISA card to sell to PC users as late as 1986.

Ditto with the loading screens from The Pawn/Guild of Thieves etc....all 512 colours on ST but on PC EGA...yuck! PCs just weren't flexible in the 80s and in the UK we were VERY picky about these things

frogstar_robot · October 31, 2010

All the ST had to do in 1985 was kill the hype coming from Apple with the original Macintosh, a terrible terrible machine.

The Tramiels were absolutely horrible at any matters concerning hype. Steve Jobs is so good at hype that he is said to generate a "Reality Distortion Field". No amount of technical excellence would have killed Apple's hype. That is just bringing a knife to a gunfight. Without changing a thing, both the ST and Amiga were VASTLY superior in technical terms to the original Macs. Greater technical refinement or better tweaking of the price/performance compromise were not was needed to compete with Job's Apple. What was needed was some P.T. Barnum inspired marketing and a bit more software elegance. Jobs comes at things from a consumer design perspective. He isn't and never was about the feature bullet list.

The original Macs were as you say "Terrible, Terrible" and very expensive at that. But they were constructed to be as discoverable as possible to non-technical oriented types and masterfully marketed. Jobs thinks very carefully about "Why would anyone want this?" and then hammers on it as hard as possible. Technical and value legerdemain alone is not enough to counter him.

Shredder11 · October 31, 2010

oky is spot on with his last post. The Archimedes looked a big deal back in 1988 but it was a luxury few could afford in the UK; even the ST or Amiga were a luxury for many of us and I for one could not afford one from my pocket money or first wages even, circa 1985 to 1990, although to be fair I was spending my money on games, music LPs(later CDs), magazines and mail order catalogue repayments, so there was nothing left to save up for an ST etc. The PC games were a joke and I can often remember laughing in disbelief, at how bad they looked in comparison with even the 8-bit games. My 8-bit gaming days fizzled out around 1990 and it was not until 1994 before I returned to computing with an 1040STE, but that was strictly used for music production only.

emkay · October 31, 2010

I don't think you understand business very well. Jack bought Atari from Warner with his PERSONAL FINANCES. Had he not done that ATARI WOULD BE DEAD IN 1984 simple as that.

Seeing the caused trouble, he better would have Atari let die.

Commodore only made business errors AFTER Jack left. He made the PET reality, he made the C64 happen with his purchase of MOS with his profits and he had no more money or MOS technologies engineers to make custom chips and spend 2 years to make Amiga-II. Everything after that EXCEPT Amiga 1000 is flawed from Commodore. Commodore 16/Plus4 = crap. Commodore 128/128D is overpriced underpowered. Amiga 2000 same shit as A1000 2.5 years later. Commodore LCD machine would have cleaned up but Commodore never released it because some idiot listened to Tandy who told him it would never sell, it was finished and complete!

Not all is as it looks for real. Commodore sold the AMIGA just in order of the success for selling the C64. Many people may have thought, the AMIGA is the successor to the C64. At Atari the AMIGA never would have reached this high count of selling.

A8 was dead, Making ANTIC/GTIA/POKEY cost Atari Warner 2-300% MORE than MOS making SID/VIC-II for Jack @ Commodore. Dead end technology, and by now the C64 coders were making awesome music on SID and getting ALL 16 colours on screen with 32 sprites multiplexed in 160x200 in 1985/86.

And?

With more effort on the A8 , we possibly would have had a "256 colour Wolfenstein 3D with real digi sounds at a playable framerate, back in 1985 (or even earlier).

Also Jack rescuing Atari from bankruptcy/shut down by Warner cost him most of his own personal money, did you expect them to design 16 bit custom chips for $10 in 6 months? Every day he owned Atari he was paying off debts racked up thanks to the rubbish sales of 5200 and A8 worldwide after Warner ran it into the ground, he made it to 1995 from nothing, second time in his business life.

A new computer was needed FAST, GEM is a genius OS superior to Mac, the ST was technically far superior to the Mac but cost 65% LESS. He did a great job in 1985 with the money he had left over rescuing Atari from the creditors. STE was a waste of time though too little too soon. Alienated the public AND the games developers and split his market and increased production costs making STE AND STFM together in factory.

What they did was clear enough. Use the knowledge they brought from Commodore to produce something like a 16Bit C16. It's not as complicated as you try to implement.

AMY was nice but probably cost more than Amiga Paula chip. Antic/GTIA were old news, no place for mono sprites and 160x200 4 colour graphics on 16bit £799 computers in 1985/86

That's why I wrote "16 Bit version" . A simple doubling of the clocking frequency and they would have fit to a PC Monitor.

+wood_jl · November 1, 2010

what could have been even cheaper than to use the hardware developed in the own house? ANTIC, GTIA, and POKEY were available. Just make them 16 bit was given.

What do you mean "just make them 16-bit?" This would be a new design, no?

But that lazy ex-commies didn't care a fart and used this YM - and this unworthy Shifter Chip.

They were going for cheap. They didn't have to design the YM, as a "16-bit Pokey" would have required. These guys didn't have any money. They went for a cheap machine, and they delivered a cheap machine. I'm not going to defend the YM but it makes enough beeps and bloops to still make an arcade game fun.

The ST series had this "slight" advantage of the straight 8MHz clocked 68K CPU, while the AMIGA had to reduce the clocking to the Chipset's needs(7.19MHz) . At the End, you couldn't even dare to compare both machines, if you're clear at mind, because the AMIGA is similar to TWO STs with some minus and some plus .....

Of course you can compare them "clear of mind." They were competitors. The ST was "color Mac at 1/3 price" and the Amiga was bells and whistles, galore. As mentioned numerous times, the games on the Amiga are better, but they're still similar enough to be compared: generally 320x200x16 vs. 320x200x32 with hardware sprites and blitter. I'm amazed some of the games on the sprite-less, blitter-less ST look and play as well as they do, though. It's not like the Amiga was running 1024 x 768 x 16.7mil colors, then I'd say we can't compare. "Arkanoid" and "Starglider" look pretty similar on each, just as "Gyruss" and "Master of the Lamps" look pretty similar on the Commodore 64 and Atari 8-bit. I know people will argue to the end of the earth that this is not true, however.

How, exactly, is the Amiga similar to two STs?

Ofcourse, compared to the 8 bit Atari games, the ST was way better in resolution and in games that needed a faster CPU. But, there often wa a sour feeling involved.

No sour feeling here. ST served well. I don't see how that's a threat to any other computer, but evidently it is, to some people.

+wood_jl · November 1, 2010

And Adlib sound is superior in every respect to the ST's onboard sound though it might not have always been used well (compare crappy ST music to crappy adlib music though...). Albeit the YM2149 did give a good route for CPU driven PCM playback while the YM3812 didn't (had they added a simple resistor DAC it would have addressed that though, but that's going to eat up CPU resource just like software playback on the ST and the SB added actual DMA audio).

Well, yeah. Sound wasn't a strong ST point even in 1985, and the Adlib came out in 1987 and took a couple more years to gain popularity. It cost $200 when it debuted (although prices fell quickly) but by the time you added the cost of the complete PC to put it in, it's an apples/oranges comparison to ST.

+wood_jl · November 1, 2010

Had the ST been sold by a company called anything but Atari the business world would have chewed your arm off to get hold of one. The 70hz 32khz non interlace 640x400 high res coupled with the DMA port for hard disks and PC DOS compatible disk format would have worked very well for a PET replacement from Commodore. Hell Commodore in the 70s was very highly respected by business customers to be honest. How things change lol

Indeed. If the ST had been branded a Commodore (or anything rather than Atari) it would have sold more. Hell, if any of the Atari computers were branded something else, they would have sold more. It's amazing how "game machine" (synonym for Atari in those days) was such a detriment to a computer in the early 1980s. In modern times, a computer called a "game machine" is assumed to be a kick-ass machine above and beyond the call of Word and Excel.

+wood_jl · November 1, 2010

That's a big difference with Europe and the UK... PC gaming pretty much followed up the wave of mass popularity of the C64 (ie '86/87 onward).

PC Gaming didn't take off until the early nineties. The mass wave of popularity of the C64 you talk about moved on to the Amiga and ST, the 16-bit generation. not PCs.

True that!

The_Laird · November 1, 2010

Indeed. If the ST had been branded a Commodore (or anything rather than Atari) it would have sold more. Hell, if any of the Atari computers were branded something else, they would have sold more.

In the US maybe. Not in the UK and Europe, Atari didn't have a bad name here. Look at how well the ST sold on these shores

Edited November 1, 2010 by The_Laird

+wood_jl · November 1, 2010

All the ST had to do in 1985 was kill the hype coming from Apple with the original Macintosh, a terrible terrible machine.

The Tramiels were absolutely horrible at any matters concerning hype. Steve Jobs is so good at hype that he is said to generate a "Reality Distortion Field". No amount of technical excellence would have killed Apple's hype. That is just bringing a knife to a gunfight. Without changing a thing, both the ST and Amiga were VASTLY superior in technical terms to the original Macs. Greater technical refinement or better tweaking of the price/performance compromise were not was needed to compete with Job's Apple. What was needed was some P.T. Barnum inspired marketing and a bit more software elegance. Jobs comes at things from a consumer design perspective. He isn't and never was about the feature bullet list.

The original Macs were as you say "Terrible, Terrible" and very expensive at that. But they were constructed to be as discoverable as possible to non-technical oriented types and masterfully marketed. Jobs thinks very carefully about "Why would anyone want this?" and then hammers on it as hard as possible. Technical and value legerdemain alone is not enough to counter him.

I didn't think Jobs was running the show back then. Wasn't it Sculley? Perhaps I am mistaken, but I didn't think Jobs worked his magic until he came back to them in later years, from NeXT.

+DarkLord · November 1, 2010

Indeed. If the ST had been branded a Commodore (or anything rather than Atari) it would have sold more. Hell, if any of the Atari computers were branded something else, they would have sold more.

In the US maybe. Not in the UK and Europe, Atari didn't have a bad name here. Look at how well the ST sold on these shores

Back to what I was talking about geographically speaking - if it was Commodore here where I live, it was almost non-existent. Our area was Atari ST, all the way.

The Usotsuki · November 1, 2010

The original Mac was during Jobs' first tenure.

kool kitty89 · November 1, 2010

I was simplifying things for sure, it was generally an issue of cost in the early VIC20/Spectrum/C64 days around 1982/83. There were probably 40 different computers for people to choose from in the UK at this time and most people looked at cost of machine and cost of games/availability of games. Atari A8 games were either on cart or disk and imported. This stacked up badly against 5.99 cassette games for Sinclair and 6.99 games for C64.

ZX Spectrum, very basic but just good enough with colour and beeper sound. £125 for 16k

C= 64 which was light years ahead in almost all aspects but cost a whopping £325/299

Acorn BBC Micro was somewhere between Spectrum and C64 but cost £600 for 32k I think.

Yep, and with Atari's position with a pretty capable machine years earlier than others, they had a chance to get into a strong position had they been able to grasp what the UK/EU markets demanded. Back before the boom started around '83 (especially with the Speccy 48k getting popular) you had a building market (users, and developers) from the super low-end ZX-80/81, low-end VIC-20, and higher-end BBC Micro plus the Atari Machines which were neither here nor there and weren't cost optimized compared to the American counterparts. (no need for heavy shielding and most of the 800's extra features were also unnecessary -short of doing a full board/case redesign, maybe they could have simply taken the 400, dropped 2 joyports, replaced the cast aluminum with plastic brackets to fit in the standard case, and added a mechanical keyboard as aftermarket kits allowed -maybe a low-end chicklet version but avoid membrane to stay ahead of the Sinclair machines -or maybe membrane for the lowest end 8k machine if they didn't just skip to 16 and 32k models only -48k was possible using the 400 board but needs modification -though for full 48k version that could have been done at the factory, let alone a later version with all 48k soldered to the board).

But by '83, having something like the 800XL in quantity at competitive (with C64 at least) prices would have been significant.

Aside from the base hardware there were also the big issues of supporting software development and tape software in general (and an odd issue of slow tape load times it seems, in spite of the Atari drives being 2x as fast as Commodore ones -short of custom loaders or replacement FSK decoders in the A8's case, but 600 baud isn't that bad, and even that could be pushed to ~900 baud with the standard FSK decoder -short of compression). But in the US it was a different matter (the low-end side was there too, but the higher end/hobbyists demanded more expandability and tapes faded quickly in favor of disks and carts alone -and with disks it did indeed have a big advantage over Commodore -Apple shared that advantage if not more so, something of a non-issue in Europe though)

Spectrum was cheap and chearful with 100s of games coming out weekly in 1983. C64 had serious graphical/sonic advantages and a real keyboard and 64kb RAM (a seriously big amount on launch day) and many games weekly by 1983. Acorn BBC had not many games, cost a fortune too but was the one your kids used at school so toffee nosed parents bought it for 'educational' reasons for their children.

In the Sega-16 discussion it came up (lookin through early 80s UK computer/game mags) that the Beebe was rather popular early on, but fell behind quickly with the surge in Speccy popularity from '83 onward (plus the C64)... the fact that Acorn screwed up with the Electron was probably a big part of that. (perhaps they should have simply produced a cost reduced, fully compatible and full performance version of the Micro instead -ie use the same integration and cost-reduced case/keyboard of the Electron, but with full speed chipset/memory -making it expandable beyond 32k would have helped too) Hell, had such a system remained popular, that would give more merit to potential A8 ports as well. (same CPU architecture and almost as fast, sound chip roughly similar, and full framebuffer bitmap modes for software rendering -especially with the 320x192 mode more practical in PAL TVs with the lesser artifacting -C64 only had the character modes, so trickier to approximate a linear framebuffer and the CPU was 1/2 the speed of the Beebe)

There wasn't much room for anyone else really, and to be honest Commodore were very lucky they had VIC-II and SID designed by engineers payed a regular salary AND they produced them at cost price. Atari could never undercut Commodore on custom hardware production costs. 64kb RAM was expensive to build into any machine in 1982 and was a massive risk.

Yep, so Atari's main advantage was their established presence on the market, but they squandered that and had already screwed up a fair bit pushing the appliance computer angle. (at leas that's what Curt and Mary seem to have surmised on the issue... that and the 400 and 800 focused heavily on being consumer level products and thus had onboard RF modulators and required hefty and costly RF shielding while the VIC-20 was produced as cheaply as possible -short of a membrane keyboard- with an aluminized cardboard RF shield and external RF modulator -not sure if Class B even exited yet, so maybe they were trying to cheat into class A, but I think it might have been Class B related -Atari had to conform to the restricting Class C with the 1979-1981 models)

For the lower-end range you did have the 16k machines which were fine for those interest in cart games. (tapes were not very popular and disk drives were still expensive, and while carts were more expensive per game, the lower cost base unit, fast loading, and simplicity were big pluses -maybe I'm missing something -conflicts of interest?- but it seems like the 400 could have been pushed handily in place of the 5200)

And it may never have fully canceled out the difference (vertical integration is pretty significant, though it did only apply to the main chipset and not a big chunk of other components and peripherals) but they could have gotten ever closer (and as it was the prices were fairly similar for the C64 and 800XL/800XE/65XE from what I understand) especially with further consolidation (ie had CGIA been introduced), though the addition of the MMU wouldn't have helped that (ie, for the low-end, maybe safer to stick to 48k, though there's a lot of factors).

OTOH there should have been other smaller chip vendors that Atari Inc could have probably purchased and cut-out the middle man, but they'd have had to weigh that with the investment risks and general alternatives to competitive outsourcing to multiple vendors. (and particularly overseas options -something I believe Tramiel tended to avoid, especially anything tied to Japan -no surprise)

Remember that most video game console manufactures also lacked the advantage of vertical integration (Sega and Nintendo outsourced to Ricoh, Yamaha, and others), though NEC had the in-house advantage there, including RAM and manufacturing the optical drive for the CD. (but in that case the hardware design itself was by a 3rd party, HudsonSoft and they squandered that advantage in the west with poor marketing) And Sony had a huge advantage not only for the chipset but owning key patents for optical drives and having the manufacturing capacity for the chips and other hardware components. (let alone the capital)

Also there were lots of tricks people started to discover even within a couple of years, the C64 char mode was a god send for colourful games using more than the 4 colours, and sprite multiplexing and then sprites in the border all added a certain glitz that none of the other 8bits had and this is pretty much why it just sold like hot cakes for a decade. The A8 was quite difficult to code and the secrets of the various quirks to exploit were much slower in arriving, by the time they did A8 was old news and ST was the Atari talk of the day in the mid 80s.

In the US that wasn't the case though, A8 tricks were rather common fare... especially since many built directly on what consoles were already doing (especially the color reloading for sprites and BG). Something that would have been very different had A8 programming been supported very well from 1981 onward, let alone if the VCS had been more popular... though Atari generally didn't want 3rd parties publishing for the VCS. (maybe if 3rd parties took the initiative like in the US, especially with something like the Starpath Supercharger breaking the cost limits of carts and allowing tapes -and more work RAM if the full 6k wasn't filled)

That would also have made it tougher to push the 7800 in those markets given some of the tricks were suited better to VCS (or A8 to some extent) experience than other things.

But I would hardly call the C64's char modes a "trick" it's a standard tile based graphics system with indexed colors like what Texas Instruments did in '79 or the VIC, or what pretty much every game console and most arcade systems did from the early/mid 80s onward (ColecoVision, SG-1000, NES, SMS, PCE, MD, SNES, and in some respect the use of indexed textures on 3D systems applies -ie 4 or 8-bit indexed graphics output to 15-bit or 24-bit RGB framebuffers, or in the Saturn's case you actually had a character/tile map for the BG).

It makes me wonder if the A8 could have done something like that had it used external CRAM instead of only the internal color registers (or a mix of both -like only using external CRAM for 2 of the entries per character and the others in -on-chip color registers), it would seem that the RIOT chip would have been a rather neat option for that (ie use it in place of the PIA and you've got similar IO, similar pin package, plus 128x8-bits of RAM -ie 128 color indexes at 8 byte each from the 16 colorx8 intensity palette). Or even with only the 9 color registers, allow some flexibility of use on character cells (maybe an offset to allow a character to have colors lie anywhere within the 9 colors selected -using the 4 sprite colors as well, or make it more granular so you only had 2 bits used to toggle the offset and have 4 different combinations of colors -though if you wanted to allow that on a per-character basis you'd need to cut the charset to 64 with 2 bits for the color selection, or have fixed groups/regions of the charset with different color sets used -in the latter case that would mean full 256 charset and for a full 3-bit offset size allowing 8 sets of 2 or 5 sets of 4 colors: set one would be colors 1-2, then 2-3, etc or 1-4, 2-5, etc) The fixed range/group palette thing could also keep 256 chars/set in the case of 128 bytes of CRAM. (in spite of only allowing 64 pairs or 32 groups of 4)

That's really idle speculation though as I don't know how much it would cost to do any of that. (but with the 1bpp character modes especially it seems a big waste to only have 2 colors usable when you have 9 color registers, and even more so with the nice broad palette available to index from compared to the limited 16 color palettes of contemporaries -with various trade-offs in selected colors, speccy probably the worst then CGA and TI99/4 -both arguable- VIC is pretty arguable too, then apple II and VICII is nice for some things but limited for some others by the drab colors -great for some shading and earth tone stuff -the computers limited to 3-bit RGB like the PC8801 and Beebe were even worse off than the Speccy but did allow more flexible use of color in-screen, and then you have the limited 9-color palette of the CoCo... and of course none of those include artifact colors in NTSC video -which helps CGA a lot and had clones commonly used color composite monitors, it probably would have been much more commonly used -but PAL users would be out of luck due to the lesser and different artifacting -let alone RGB users)

In the case of the TMS9918 release the same year they had the color indexes in video DRAM, not dedicated CRAM or color registers (other than sprites), so a somewhat different route not taken by others. (in the case of mode 1 you have color indexes set for groups of 8 characters, so only 32 indexes or 16 2 color pairs selected from the 16 color master palette -mode 2 added to the TMS9928 expanded that dramatically to 2 colors per 8 pixel row and thus 8 index pairs per character and 2 kB used for just the color indexes and limiting it's use mainly to splash screens -perhaps a shame that there wasn't middle ground with a mode allowing one index pair per single 8x8 character -ie only 256 bytes and still far more flexible than mode 1 and more like what the Speccy's attributes allow)

Ahh when I talk about custom chips I mean not using the CPU to scroll the screen, move graphics memory, create moving graphics like Sprites or Blitter Objects on C64 and Amiga respectively. As I said Commodore had MOS make VIC-II/SID and then bought Amiga computers and again made the chipset themselves. Although they paid a lot more to get the Amiga chipset than it cost them to end up with the C64.

well that's a rather inaccurate and specific definition you've got there. And you can argue what "cpu to scroll screen" even refers to as you have character based displays where you simply remap the characters (names table, tile map, or whatever term is used) and you have simple cell-wide scrolling and object manipulation.

TIA in the VCS is obviously a custom chip... And the sound chips would also be among those. Custom chips DO become off the shelf parts once they're offered to the open market. (the SID eventually became an off the shelf part and again, the TMS VDP and SN76489 both became off the shelf by the early 80s)

And you DID have off the shelf coprocessors for blitting and such, though they weren't cheap (and more often entailed use of DSPs or -for lower-end stuff- CPUs/MCUs dedicated to graphics acceleration). You had hardware scrolling in VGA and some things to aid software blitting, then you had real dedicated graphics acceleration coprocessors appearing as well (not jsut off the shelf DSPs but GUI accelerators and bit-blit engines), especially by the early 90s. (but games didn't really start to push use of hardware acceleration until 3D from '95/96 onward, especially with Win9x -though in many cases CPU rendering was still supported as well up through the late 90s) I think the earliest example of a more fixed-purpose accelerator was IBM's 8514.

From this point of view the ST didn't have anything bespoke/unique to it. There was no scrolling/animation in hardware and the sound chip was more or less identical to that from a lowly Spectrum 128k or Amstrad CPC 8bit computer as far as consumers were concerned.

It did have the attribute cells and a linear framebuffer, which is more than can be said for some systems. (to the extent that you could manage smoother software scaling than character only systems and more than 2 colors for a 1bpp display)

The Timex computers were not significantly cheaper than the Commodore 64 in the USA? Sure software would be a problem and require importing (so reversing the situation the A8 was hindered by in the UK in the very early 80s). Personally Americans had far more disposable income, the Spectrum was unique to the cash starved Brits who WANTED cheap and cheerful so doubt it would make a difference, the Spectrum just has too many compromises to be worth the few quid you save.....unless you don't have a few quid extra and the alternative is nothing at all to play with or wait 2 years and buy a second hand computer etc

The Timex machines were flash in the pan hits that sold because of the price but went nowhere for most users (the so-called doorstop syndrome even more extreme than with the VIC-20 -which suffered from that more because of the short lifespan and shift to C64 than general lack of support). And the US didn't have the same sort of strong homebrew programming interest as in Europe, especially not at the age ranges in Europe.

And the Spectrum was never released in the US proper and what we did get may have fit the market better (keyboard more like the Speccy plus, built-in joystick and cart slot, built-in AY8910 2 years before the 128k), but it ended up without a significant price advantage over the C64 (let alone other machines on the market dropping to compete or being dumped) and had an incompatible BIOS ROM. (simply fixed by a cart with a Speccy ROM, but that was only available from homebrew/3rd party).

Even if Timex had brought the Spectrum 48k over directly (more or less as they had the ZX81 and such), it would have had to been modified and be more expensive due to the mandated RF shielding. (so larger case, more cost, etc) There was the 2048 which was never released in the US, but was much closer to the Speccy (only main addition was a couple graphics modes with 1x8 attributes and 512x192 highres monochrome mode) and a closer form factor the the 48k.

They did release the Timex 1500 which was an upgraded Timxex-1000/ZX-81 with 16k at $79 in 1983 upgradeable to 64k, but the same limited CPU/Graphics of the older machines. http://oldcomputers.net/ts1500.html (neat silver incarnation of the 48k case design though)

The Speccy's fast tape loading speeds might have been significant in retaining interest in the declining media, but that didn't help the CoCo... (albeit the CoCo wasn't as cheap as the Speccy might have been) Importing/localizing EU software would have been critical in any case.

All of the mass market US machines were significantly cheaper than EU/UK contemporaries (presumably due to vatious taxes), so it's hard to tell exactly how some other might have been priced.

It's rare to see a Megadrive game which doesn't have 64 colours, isn't using full screen parallax and does not use all the features of the soundchip with FM AND AM sound routines for music and speech/samples etc. Compare Gauntlet 4 from Megadrive to Gauntlet 2 on Amiga. The Amiga game looks same as the ST version, the ST version is INFERIOR to Gauntlet 1 in all but horizontal scrolling (screen size smaller, tiles are 2 colour, sprites smaller, animation slower, less speech/samples etc). If you look at Gauntlet 1 on ST and compare to Gauntlet 4 on Megadrive you get an accurate gauge of the difference in power of both. Now run Gauntlet 2 Amiga side by side with Gauntlet 4 and a n00b on the street would think the Amiga is graphically inferior to an ST running Gauntlet 1!

Nope, the majority of MD games have significantly fewer than 64 colors just like most SMS games have closer to 16 colors. Aside from a few games using shadow/HL or raster tricks (palette reloading) you're pretty limited in the on-screen color count if you want good, smooth graphics without clash. You have a maximum of 61 colors possible normally: 4x 15 color palettes and 1 color BG color and inevitably you have a lot of redundant values in those palettes (especially when you're restricted to 9-bit RGB -more so with the SMS's 6-bit RGB and 2 palettes -only 1 available for sprites). It's the same reason you rarely see even more than 100 colors on-screen with the SNES or much less than 25 colors on the NES. (4 3-color sprite palettes, 4 3-color BG palettes, 1 solid BG color) With the PC Engine you have 512 total color index but even then you're not going to practically get anywhere near that and would be more likely pushing fewer total colors than the SNES as you're limited to 9-bit RGB (though more likely more colors per tile on average due to greater potential for optimization) and it's arranged with 16 15 color sprite and 16 15-color BG palettes and 1 solod BG color. (ie 481 colors without added tricks, far less in practical use due to redundancy)

Same reason the Neo Geo rarely has even remotely close to even 1000 colors on screen and often more like 200-300 colors, still only 15 colors per sprite tile (only sprites in that system), but with 256 15 color indexes selected from 32,768 colors (2 banks with 2 intensities, but only one used per screen aside from raster interrupts) and thus theoretically up to 3841 colors but far far less in practical use. (SMS, PCE, MD, SNES, and many contemporary arcade machines all used 4-bit graphics with indexed color palettes)

The SNES does have 8bpp modes but they were never used for 2D games aside from mode 7. (they use more ROM, more DMA bandwidth and 2x the VRAM space, and you still wouldn't have 256 colors aside from mode 7 as the 256 indexes are shared in the sprite/tile indexes and you will thus have all the transparent values limiting things to 241 colors max, but more so if you want any decent use of 15-color tiles or sprites as you need to optimize palettes further and inevitably have more redundant values -minimized with careful opimization)

So in many cases using 32 colors (31 colors rather) on the Amiga using 5 bitplanes would allow for better looking graphics than the MD. (larger palette, no restrictions of fixed palettes per cell, etc) Or even some advantages over the SNES (short of 8bpp splash screens) Now had the SNES supported 5 or 6 bitplane modes that likely would have been far more useful than the 8 bitplane mode. (Genesis uses 4-bit paced pixels, so not an option -and the shadow function has to waste a full 4 bpp plane just to do something like the Amiga does with EHB, granted it also has shadow which the Amiga doesn't)

Though using the 7/8-colors per plane dual playfield modes would limited the Amiga's color a lot more compared to the MD. (even more so if 4-color sprites were used) But in cases where the 31 color mode would be practical, the Amiga would more than likely have an advantage. (maybe not in full colors on screen, but far greater flexibility of using those colors anywhere on screen, and the Amiga has fast palette tricks with the copper while consoles had to do that with CPU resource and the Genesis has nasty artifacts on the line after that's done -usually masked by flickering sprites and limited to water effects)

Highlight is limited (only used with sprite planes onto other sprites or BG), but shadow can be activated by overlaying a BG layer: so technically you could pair 2 BG layers and with carefully selected colors (no enteries that are exact 1/2 intenssities of others in the same palette) you'd have up to 46 unique color/shades per 8x8 tile (16 colors with 2 shades and the remaining 14 colors of the sacrificed plane), but only 1 scrolling BG and having effectively 8bpp graphics (double the memory used), but you would still have 4 subpalettes as well, so multiple sets of up to 46 colors, and 15 color sprites. That was not ever used for any games AFIK and likely for the same reason the SNES didn't use the 8bpp modes (shadow and HL are used to limited extents for shading effects in many games), though it might have been more feasible to only do that on a few rows of the BG and not others. (meaning not only less memory used, but normal 2 plane BGs for the non paired rows) Again, I don't think that was done either.

However HL/Shadow were used to substantial extents for still images just as the SNES's 256 color mode was at times (and Amiga HAM beats them both in many cases, or even EHB+copper tricks), toy story is one such case and a handful of FMV took advantage of that as well, but the only one I know of is Sherlock Holmes 1 and 2. (and it's uncompressed I think with very careful use to maintain an acceptable bitrate -not used in any of the compressed video codecs intrestingly enough, even though that could have helped things considerably... actually HAM on the Amiga likely would have made a good compression scheme alone for the CDTV -let alone with some additional lossless/lossy algorithms applied or HAM-8 on the CD32, as did the plain color inexing to 4bpp tiles of the PCE and Genesis -far more flexible on the PCE but still usually only about 40-60 colors when clash is avoided, or ~30 on the MD, but Night Trap managed close to 48 without other tricks but plenty of clash... the SNES actually has a 2bpp tile mode with 32 subpalettes broken up that might have served as a good simple compression screen as well beyond the 4bpp modes -or doing things with the PCE or SNES like dropping bitplanes on a per-tile basis as a simple form of lossy compression -which many early games use to "compress" graphics in ROM and then pad to 4 planes)

But now I'm way off topic.

Back to the quality of games... it's debatable but even most Sega fans (or especially Sega fans) remark on the vast range of quality on the MD... from some amazing music and optimized graphics to crappy music even by Adlib standards, terrible PCM playback, sloppy graphics, etc. And more arguing general issues of art design and taste, or simply game quality. (some tend to beat down on the US developed stuff in particular)

It's not really different from the NES, SNES, PSX, or PS2. (some would argue there's more MD shovelware than on the SNES)

Sadly very few games on ST/Amiga are anything other than the bare minimum, and total rush jobs by cheap ass UK Software houses spending the cash on arcade licenses and then hiring rubbish talentless development groups to hash it all up. On SNES/Megadrive/PC-Engine AKA Turbo Grafix the norm is very polished code and lovely graphics and sound from talented people. Gotta love those industrious Japanese games coders!

You saw the lower-budget/rushed stuff more often on computers in general (including PC), but consoles required more commitment due to the expenses of media and licensing agreements, but even so you had some sloppy/poor stuff including ports of some already mediocre computer games or sloppy ports or better computer games.

But if you're purely talking about Japanese developed stuff, then yeah, there's much less crap on average and more mediocre than downright poor games, but again I was thinking much more toward the western developed stuff. (some of Acclaim/Activision/Accolade/EA/or even Sega published stuff -in some cases selling quite well but criticized)

Even in early 93/late 92 a 486SX 25mhz with no sound and the slowest ISA SVGA card cost £1000+ from the most dodgy clone makers. In the UK around sometime between Amiga 500 and 1200 Amstrad tried to launch the Sinclair PC10 which was an all-in-one style 8086 CGA machine with 512kb/360kb drive and blurry TV modulator to compete at around 400-500 quid. It failed, the worst computer format launch in the history of computers. Nobody wanted CGA in the time of C64 or EGA in the time of Amiga. Shadow of the Beast was impossible on PC technically before 1990 and even then in 1992 a machine to play Shadow of the Beast would cost £1500+ Ditto Super Stardust in 1995...A1200 £299 Pentium 120mhz £1000+ Ditto Lotus II Turbo with A600 costing £299 486DX2-66 costing + Roland sound card AND Soundblaster = £1500+

You're talking pre-built and higher-end stuff... by the early 90s the best options by far was buying off the shelf and/or used components to make the most cost-effective machine possible (or buying a used machine with few features but lots of upgrade potential and going from there), or in the case of those without tech expertise, paying someone else to build such a lower-end custom machine is also a better option, much more so if you have a friend or family member who could help out.

No reason to go to MT32... that was a high-end exception. Best options by the early 90s were Soundblaster/clones or similar ones like the Pro Audio Spectrum prior to the Ultrasound (and you really wanted a SB compatible for full range of adlib/SB support). One problem was that Creative cards tended to have some pretty crappy analog circuitry meaning distorted/muffled output though the SBPro/16 were more acceptable. (some of Yamaha's compatible offerings were the best though and the Pro Audio Spectrum was above average, while lower-end clones were worse)

With the original 1989 Soundblaster you got the OPL2 synthesizer of Adlib and a single 8-bit DMA audio channel with up to 23 kHz sampling (and ADPCM decoding support) while the 2.0 boosted that to 44 kHz, then 22 kHz 2-channel/stereo for the Pro and dual OPL2 chips for 18 channels and stereo, the Pro 2.0 switched to the much more capable OPL3 (18 channels with L/C/R panning per channel, added operator waveforms, and some limited 4-op FM modes -additive synthesis and use of the alternate waveform are often more useful), and the SB-16 added 16-bit 2-channel/stereo at 44 kHz. (in all cases the DMA audio facilitated software mixed MOD players and sfx, but it wasn't until about 1993/94 that MOD music became common and synth stuff was far more common... unfortunately the OPL3 was never fully taken advantage of and the DMA channels were rarely used for music and almost never were both used for music and SFX -ie like the MD did with PCM drums and SFX as well as FM sfx and music)

General MIDI and other sample based cards pushed in before the SB-16 really got pushed and CD-DA or compressed streaming audio (often in WAV format -often lower sample rate/mono PCM or ADPCM) appeared concurrently.

So for the very early 90s (~1992), the best low-cost PC would probably be VGA, SB2.0, 640kB-2MB, a modest HDD (maybe 20-60 MB), 5.25" and/or 3.5" FDD (preferably HD -and most games were available on 5.25 into the mid 90s aside from CD exclusives) and a fast 286 (ie 16-25 MHz) or maybe a 386SX. (in the latter case you have potential for some pretty significant upgrades via Cyrix's 486SLC line -even with clip-on chips for surface mounted CPUs) Preferably with a used case and color VGA monitor. For a more mid-range machine for '92-94, probably a 386DX 2-4 MB (4-8 MB by '94/95), VGA card, SB2 or possibly 16, 2x CD-ROM drive, larger HDD, etc. There's a lot of contextual stuff and buying pre-built machines was usually a bad option... This came up on Sega-16 too: http://sega-16.com/forum/showthread.php?t=13161&highlight=costs+ages

But one big part of it was sheer availability with more PC vendors and more software (especially by the late 80s) in general, or at least easier to find software due to broader distribution. (and lots of potential for upgrades with different limits depending on the motherboard and CPU, but even the 386SX had some decent option, plenty for 386DX -especially with FPU support- and lots for the 486 and it goes on with 3rd parties offering faster CPUs for older motherboards/sockets at lower prices, especially Cyrix and AMD, though 486SX was a waste and a 386DX board with provisions for FPU would be preferable by far -486 accelerators for 386 sockets and 387 support rather than the ripoff 487 -FPUs became important for games by about 1995, especially with the release of Quake)

The point is Commodore never ever did anything useful with the Amiga chipset and refused to licence it to Amstrad either. The A500 was as close as they came to building a spiritual successor to the C64 but technology was moving at an ever faster pace. CGA to EGA was longer than EGA to MCGA/VGA and it all snowballed. Ditto 8088/86/V30 to 80286 affordable PCs to the later years of breathtaking pace of ever increasing Mhz of 486/Pentium on a monthly basis! Commodore were trying to compete with 25mhz CPUs and Atari with 16mhz CPUs in Falcon with 16bit RAM not 32bit! People gave up anyway by the time Commodore went tits up. People got fed up with a slow CPU. If you can't make an awesome custom chip to run DOOM on a 14mhz computer then give up on custom chip development and ramp up the CPU speed to match the brute processing power of competing machines (ie in 1993 that would be 386DX 25/33 competing with 14mhz crippled 68020 with no 32bit fast ram included off the shelf...madness you get 200% CPU speed with just 512kb of non chip ram on an A1200...the machine is effectively running at the speed of an 8mhz 020 as sold by Commodore!)

Yeah, aside from licensing and potential for a boarder standard, they missed good opportunities to upgrade the base chipset much earlier than they did (I seem to recall a faster blitter and 256 color/8bitplane support much earlier that was passed on... even if it was still 12-bit RGB, much more than what ECS did... though upgrading Paula to 8+ channels and hard panning per channel would be very significant -let alone smooth panning)

If they were going to go with faster CPUs they needed to do it much sooner anyway and short of other custom chips, 8-bit packed pixel support was a big deal too (especially if the blitter was enhanced to work especially fast with that -even potential to buffering for 16 or 32-bit writes... or 64-bit even).

You weren't going to get accelleration for Doom unless you went for a fast DSP or DSP-like chip that was suited to ray casting and vertical line rendering. (traditional 3D accelerators/blitter with affine texture rendering or line based operations would be of very limited use aside from scaled opbjects, hence why the game was totally rewritten for the PSX to use a polygon renderer)

So the Falcon's DSP could indeed have given a lot of potential for such, but before that there were other 3rd party options as well with various graphics-oriented DSP-like chips from TI and such, or CBM could have invested in developing such a custom chip themselves. (Atari did just that with the Jaguar, but that was way too late to be applied to the ST line and it was far more than just the RISC GPU, but a powerful object processor and blitter -doom was pretty much all GPU though, with blitter used for the smooth shading) But from that standpoint, in-house should have been about getting the system at least up to 256 colro packed pixel support, some upgrading to the blitter (maybe even scaling or rotation/affine line rendering) and some upgrade to paula, and then push for a flexible GPU or a lower-cost 3rd party DSP/GPU, or simply faster CPUs. (ray-casting and other non-polygonal rendering methods were very significant up to about 1996 after which pure polygonal stuff really got big, so if they didn't bother with flexibility for height maps beyond fast CPUs, they could have pushed for more polygonal oriented stuff for the blitter like affine line rendering, smooth shading once highcolor was supported, and eventually polygon rasterization -though a flexible DSP-like chip or such would be useful for all of that and fast 3D vertex calculations on top of that -otherwise the latter would fall to the CPU)

You needed something close enough in programmability to stand in for a CPU, and apart from DSPs or the like (or custom GPUs of a similar sort), there were low-cost RISC CPUs... and short of switching architectures altogether they could have used them as accelerators. (there were PPC accelerator cards, but ARM CPUs would have fit far more with the low-cost side of things)

But a very real problem with Motorola CPUs themselves was keeping up with x86, especiall after Motorola severely cut open licensing with the 020 onward vs the rather well distributed 68000, but that shouldn't have stopped Atari/CBM from bumping up to 10/12/16 and eventually 20 MHz 68ks (16 by the late 80s though). Broad success of 68k based platforms would increase potential for continuation of the architecture, so that's a big integral what-if. (especially if Motorola opened up licensing again soon enough to facilitate 3rd party enhancements -Hitachi had been a big one and as it is managed some better 68k derivatives than Motorola -their 68HC000s tend to consume a good bit less power than motorola's)

Hell, even without licenses for later CPUs, 3rd parties could have taken it on themselves to build on the architecture with enhanced compatible CPUs or even just faster and faster 68000s with more address pins (and probably added instructions, like those of the 010/EC000) and perhaps expanding the prefetch (short of actual caching) and/or an on-chip MMU and later FPU. (aside from prefetch and the added instructions, most of that is just consolidation and higher clock rates) With the 68k's small die and newer and newer chip processes that should even have allowed some pretty hefty clock speeds before heat sinks were needed (more still for CPU mounted fans).

Actually it's a bit odd that in all these years, plain 68k/MCU dirivatives didn't push into much higher clock rates. (ie hundreds of MHz, or anything beyond 20 MHz for that matter) Hell, that could have helped a lot with the Jaguar too. (faster 68k, still low-end option cheaper than an EC020 or various Intel or RISC CPUs, but a lot better than the 13.3 MHz speed used, 26.6 could have helped a lot, but given the bus speed and use of fast-page accesses, a 40 MHz 68k should have had no waitstates, but at that speed in 1993, probably pushing more expensive)

In the UK we never had such machines really, and most crappy 8086 'affordable' PCs had 640k RAM, 5.25" disks from C64/A8 days and CGA graphics in crap looking cases. Amstrad probably made the best of the worst. Their answer to the ST/Amiga was the Sinclair PC10...see this monstrosity here

It's actually rather odd that Tandy didn't bring it over or license it like what happened with the CoCo (Dragon 32 and such), especially with the more price-sensitive market. http://home.comcast.net/~rcmerritt/gaminghistory/gaminghistory1986.htm

The price was close to an ST with an RGB monitor and the performance was certainly less (for base models at least), but PC compatibility and expandability tied to it. (and that page actually downplays the sound as it's a SN76489 like the CV/SMS with 3 square wave 1 noise channel plus the normal PC beeper -usually used for SFX)

And the specific enhanced graphics and I think packed pixel. Actually back in 1982 there was the http://en.wikipedia.org/wiki/Plantronics_Colorplus offering nearly identical modes as the PCJr and Tandy later did, but it never gained enough support to be worthwhile beyond CGA support. (lack of common drivers and BIOS support killed it -apparently they didn't offer replacement BIOS ROMs either) But that wouldn't have addressed sound and that's something developers latched onto pretty quickly regardless of being standard or not. Actually a little odd tandy didn't offer standalone audio/video upgrade cards and replacement BIOS ROMs to meet Tandy standards. (might have been enough to push those as true de-facto standards and garner backwards compatibility, at least in later 3rd party cards) And it it was indeed packed pixel (especially if not using the dumb CGA interlace) it likely would have gotten better support and higher performance than EGA for games. (packed vs planar)

Ad again, with PAL you didn't even have the option for the NTSC artifact color "modes" . (and with composite you still got niche sharp grayscale modes with colorburst disabled)

I have always confessed that had Atari opted to replace the original ST with an 8mhz/16mhz switchable CPU ST rather than the 8mhz STE with inferior [to the Amiga] updates that nobody wanted to rewrite 1000s of lines of code for things would have been different up until around 1990/91. 91 was when there were VGA versions of arcade/action PC games not just EGA. I don't mean simulation stuff from USA I mean conversions of games like Lotus III etc. 16mhz 68000 Instead of wasting all that money on the STE enhancements that were never really used in exclusive games and isolated the ST/STM/STF/SFTM owners AND split the market for the software houses who had to decide what game to do and which format and then which bloody model of ST/STE etc. Just play games on STeem emulator at 16mhz and you can have a lot of fun. Gauntlet 1 scrolls better and is awesome and Lotus II is approaching Amiga levels of smoothness.

And that's probably the same reason PC games didn't support acceleration prior to high-level driver/API support (especially in windows) became common. (3D happened to coincide with that )

The sound upgrade was needed and it had trade-offs (2 channels with panning and up to 50 kHz vs hardwired stereo pairs of the amiga with ~29 kHz max at SDTV sync rates)

But yeah, not just CPU over the blitter (which would also avoid the delayed blitter issue with the MEGA), but rather than competing with the Amiga's hardware acceleration, push to one-up it in other areas like upgrading the shifter to something closer to VGA (maybe not all the features of VGA or the TT Shifter and maybe not even hardware scrolling, but at least 12-bit RGB and a 320x200 8-bit packed pixel mode with 256 indexed colors -in which case many software rendered games would perform better than with the planar modes AND have much more color than anything but VGA or the Archimedes -obviosuly important for graphic design stuff as well... if they added genlocking, they could even cut-in on the Amiga's editing/effects market and unlike PCs have SDTV compatible sync rates -albeit CGA and EGA did support that but a lack of genlocking made it moot and the dropped support with VGA finalized it until many years later with things like ATi's all in wonder card) So main focus on upgrading the shifter, no blitter, and pushing for the DMA audio sooner.

Obviously had Commodore or Atari sold as well as their first big hits (C64 and VCS respectively) then ploughed the money back into a next gen update sooner then things would not have gone the way of 14mhz A1200, 16mhz Falcon or 66mhz PC as your only choices!

And a big problem was the upgrades were simply too late, not just insubstantial. (a 16 MHz otherwise vanilla 520/1040 ST by '86/87 alone would have been very substantial, let alone making the MEGA more expandable -16 MHz CPU, general purpose expansion slots -maybe even ISA compatible- optional FPU, etc, and then the shifter and sound upgrades -you had the off the shelf YM2203, but that wouldn't solve the lack of DMA audio -but rather significant nonetheless and the 16 MHz upgrade would make software drive pseudo DMA via the square wave channels more practical)

All the ST had to do in 1985 was kill the hype coming from Apple with the original Macintosh, a terrible terrible machine. I think the styling Atari chose for the ST was a mistake in 1985 in some ways. It wasn't cheap enough at £799 inc Mono monitor to be bought by people who would be alienated by a nice sexy 3 box design like the original Megas and it was looking too much like an 8 bit computer their children played games on than a serious tool for someone who would buy a £799 machine with all the performance of a £2500 Macintosh and appreciate GEM, DMA hard disks, excellent serious software potential.

Too bad Atari didn't have the funding to really push big time marketing line that (especially television)... but in the case of the expansion issue: again, with practically no added cost a general purpose expansion port in place of the cart port would have been a big step up. (probably need 64 pins to really be comprehensive, but that would still only put it at about 8-bit ISA size )

They probably should have launched the 520STM in parallel with the 520ST! It's not like it is any different to the original ST and modulators cost peanuts so that would have allowed the early adopter to get right in there in 1985 with a shiny new computer at something like £150 less, home computer owners didn't want to pay £999 for 520ST+SC1224 and the mono monitor is useless to someone trading up from an old 8bit colour computer. SM124+ST was only really going to work for Macintosh comparisons. To be honest the ideal perfect launch of the ST would be the 520STM to home computer elite users (like the ones who bought a BBC Micro for £800 with disk drive in 1982) and a 512kb machine in the style of the Mega ST to break into serious/business user market. Sell the 520STM @ £550 in 1985 with SF354 bundle (or actually launch some bloody software on cartridge!) and then sell the 'Mega 0.5 ST' with SM124 and SF354 for £799ish to people who liked the Mac but weren't stupid enough to pay £2500 for a 128k monochrome piece of crap.

Do the STFM, vanilla ST/STF (with C-sync and SCART cables for EU), and MEGA (with multiple external expansion slots and maybe a faster CPU -or socketed CPU and provisions for a 16 MHz toggle mode and maybe FPU socket) would have made even more of a difference and none of that custom hardware. (but the latter meant an additional PCB design) MEGA by '86 would still be significant. (especially if it could be 16 MHz... maybe 12 MHz if that didn't complicate clock toggling/dividing)

I have always suspected that the Atari ST that Shiraz Shivji designed was based on plans for a 16bit successor to the PET before he left Commodore. If you look at the ST in this light you will see it fits really well. It has the right mix of colour and resolution for the time to help with doing business graphics as a successor to the PET would need. Needs to be easy and fuss free to use because business users are not loyal and don't appreciate difficult to use systems so GEM would have been perfect. Make sure you can swap data with PC disk format check. Make the machine faster than comparable computers of the time with a clean fast 8mhz 68000 design with 4mb addressable RAM. Include excellent hard disk interface on board with DMA port. Sprinkle some unique features like 3 channel sound and 512 colour palette too. You see how it makes sense? Had the ST been sold by a company called anything but Atari the business world would have chewed your arm off to get hold of one. The 70hz 32khz non interlace 640x400 high res coupled with the DMA port for hard disks and PC DOS compatible disk format would have worked very well for a PET replacement from Commodore. Hell Commodore in the 70s was very highly respected by business customers to be honest. How things change lol

Built in monitor like the PET, and you're even more like a color MAC.

But I'm not really sure it had any relation to CBM projects, maybe concepts, but probably not any solid development. It's mainly just the shifter, GLU for interfacing the off the shelf chips, and simple MMU chip for mapping memory, the rest is plain off the shelf and from what I understand most was developed in a very short time in '83/84 (they also did the custom port of CP/M as CP/M 68k was less desirable iirc).

I'd imagine Curt or Marty would know more on this already, if not interviewed Shiraz personally. (is Shiraz still alive?)

PS I should point out, in case anyone doesn't know, the original 520ST or 520STM for me is the most stylish all in one computer ever made design wise, and my comments are merely the perception this form factor gives to potential business users of your machine who all have used computer terminals hooked to mainframes, Apple Mac/Lisa or IBM compatibles and who only see this type of all in one design on 'kids computers' they purchased with 8bit technology for their kids to play space invader clones on a few years ago

I think I like the look of the XE consoles better actually (simpler and sleeker... I know Curt hates it though compared to the XLs ), but I believe all have the some ergonomic issues with too angular key caps plus the mushy feel from the springless dome switch construction. (moden keyboards tend to all use dome switches too, but usually with scissor spring mechanisms, sometimes plastic... though not more recent Apple stuff -the mushy macbook boards were bad enough but then they took those and used them for the imacs and MacPros when they'd have some pretty nice full-throw keyboards back around 2004/05... and don't even start on the mice)

Edited November 1, 2010 by kool kitty89

kool kitty89 · November 1, 2010

I don't think you understand business very well. Jack bought Atari from Warner with his PERSONAL FINANCES. Had he not done that ATARI WOULD BE DEAD IN 1984 simple as that.

Nope, not at all, Atari very likely would have been better off, but that's not Jack's fault, but Warner's for chickening out and splitting the company up and doing so in a very sloppy manner that ensured that there was plenty of confusion during the transition to TTL (which was renamed Atari Corp), James Morgan (president of Atari Inc) wasn't even notified of the pending sale until the last minute just before the documents were signed and on top of that the deal went through over 4th of July weekend, confusing hapless staff even more, it was a mess. (again, not Tramiel's fault though)

Had Warner let Morgan do his job things could have been amazing for Atari Inc: in 1983 Morgan was brought in to replace Kassar and start much needed reform of the company (which really needed it a year earlier) and by mid 1984 his plans were coming along nicely with every indication of rebuilding Atari into a lean, mean competitive company along with NATCO. The 7800 was primed for launch after a surprisingly positive reception (for the market at the time) earlier in '84, Atari had various advanced 16-bit chipsets shelved with tons of potential (enough to hand the Amiga its ass and then some given what Curt and Marty have indicated), the XL 8-bit line pushing ahead with the much needed 1090XL expansion system on the way, the 2600 Jr in the works, the 5200 planned to get a 7800 adapter to bridge the gap (not sure if the 5100/Jr was still in the works, but the corrected joystick was), and then there's the Amiga based console planned for that fall as well with a prototype PCM all ready for testing and just waiting for Amiga Inc to deliver the chips... and those were to be delivered about a day before the Tramiel/Warner sale deal started (chips were to be delivered September 30 iirc), but Amiga was at the time pushing to partner with CBM and was trying to worm out of their contract with Atari and outright lied saying they failed to produce working LSI chips. (and returned Atari's 50,000 investment with interest in attempt to cancel the contract) And interesting to note that Tramiel successfully negotiated the Amiga contract from Warner as part of the sale but they contended the 7800 until he finally paid for MARIA and GCC's 10 launch games in early 1985. Of course Amiga never got held to their contract and never delivered the chips (giving nice grounds for a lawsuit -and a good counter to CBM's suit over the ST), albeit even had Tramiel gotten the chips they'd have been stuck to the deal for 1 year purely as a console then allowing a keyboard/computer add-on but still limited to 128k (aside from 3rd party expansion, but none 1st party) and finally allowing the full computer to be released without restriction, so the RBP/ST likely would have gone ahead anyway, but perhaps provisions would have been made to facilitate the Amiga chips for later upgrades (by '87 the entire chipset could have been used).

Again, Tramiel did already have the advanced technology division's 16-bit protoypes and some engineers did attempt to notify him about those it seems, but he was focused on the ST and in some cases some key staff had already been lost (that severely complicated use of AMY -which was part of the previously planned 16-bit machines). Granted it's also unclear if any of those designs would have fit the mass-market niche Tramiel was aiming for (several were high-end workstations, though that was back in the context of late 1983), but I'd imagine there was plenty of potential for implementing those chipsets (or parts of them) into a lower-cost machine and there were several designs (the high-end workstation with 2 and later 3 68000s planned a single 68000 machine with the Gaza and Sierra 68k designs and 2 or 3 distinct coprocessor chipsets with silver/gold/rainbow). And if those were only wirewrapped protos (no LSI silicon, just TTL), that would also have been a factor in Tramiel turning them down for the sake of time. (would have been interesting if Shiraz had inspected them and perhaps made provisions in the ST design to facilitate later integration of some of the Atari Inc hardware) Plus they were developing a corresponding BDS Unix derived OS with GUI for those 16-bit machines.

But the Rainbow chipset was at least low-cost enough to make Atari consider it for a game console it seems. (Marty commented on that but he wasn't definitive) In hindsight, it probably would have been a good deal better for Atari to focus a bit more on getting at least 1 of their in-house 16-bit chipsets up and running (likely the least complex/expensive)... and that also would have made it far more attractive to Tramiel.

I assume you've seen this already: http://www.atariage.com/forums/topic/152112-sneak-peak-amiga-atari-design/

And some of it came up again here: http://www.atariage.com/forums/topic/166993-atari-7800-forum-description/page__st__75__p__2123084#entry2123084

So had Warner stayed the course (and refrained from interfering with Atari operations -which there were still problems with), Atari Inc very likely would have had the 2600 Jr (low-end), 7800 (mid-range), and Amiga based console as the high-end machine.

Commodore only made business errors AFTER Jack left. He made the PET reality, he made the C64 happen with his purchase of MOS with his profits and he had no more money or MOS technologies engineers to make custom chips and spend 2 years to make Amiga-II. Everything after that EXCEPT Amiga 1000 is flawed from Commodore. Commodore 16/Plus4 = crap. Commodore 128/128D is overpriced underpowered. Amiga 2000 same shit as A1000 2.5 years later. Commodore LCD machine would have cleaned up but Commodore never released it because some idiot listened to Tandy who told him it would never sell, it was finished and complete!

Yes, and the same seemed to happen at Atari Corp... they entered a downward spiral as soon as Jack retired in '89 (or was it late '88?), though Mike Katz left at about the same time for that matter so they lost one of their main strengths in the entertainment division as well. (ie Katz was gone by the time the Lynx launched... actually he was at Sega about that time laying the groundwork for the Genesis's later massive success in the US -he was replaced by Tom Kalinske in late 1990/91 -there was a transitional period iirc)

But yeah, a bunch of odd mistakes: dragging their feet getting a lower-end console form factor Amiga out, getting proper successors to the A1000 out (A2000 was more expandable but nothing more out of the box than a bulky, expensive A1000), then the somewhat odd and inefficient C128 alongside the Amiga, the odd and unnecessary Plus/4 and C16, the mess that was the C64GS (at least Atari had a rather usable and attractive machine with the XEGS, and back in '87 when a C64 based console could still have mattered -though had it not been half assed and had it been released in '85/86 a C64GS might have worked OK... but pushing for an Amiga game console would have been a much bigger deal, especially if out by '88). The C65 was interesting, but by the time it was planned for release it already didn't make that much sense... OTOH had it been in place of the C128, that could have been amazing (a bit cut-back of course, but same sort of idea of an Amiga/C64 hybrid for the lower-end market -especially since tacking C64 compatibility on to the Amiga wasn't really a cost-effective idea though the 6502 would have given some nice added resource for software mixing on Paula among other things... or genlocking with the VICII... Hmm maybe that wouldn't have been such a bad idea either and had they done that it would have pushed the Amiga a bit higher-end still but they'd have the C65 to address the mid-range and with the amiga having both chipsets in their entirety it could be C65 as well as C64 compatible not to mention potential for later cost reduction via integration. -hell probably managing integration to benefit the whole range of products like merging the SID and VIC II to make the C64 cheaper as well)

A8 was dead, Making ANTIC/GTIA/POKEY cost Atari Warner 2-300% MORE than MOS making SID/VIC-II for Jack @ Commodore. Dead end technology, and by now the C64 coders were making awesome music on SID and getting ALL 16 colours on screen with 32 sprites multiplexed in 160x200 in 1985/86.

A8 wasn't dead until the end of the 80s, but by '84 they'd lost their chance to really compete with the C64 directly... it was profitable enough to merit the XE redesign and selling it through the late 80s. (albeit only in the hundreds of thousands per year, not millions and tapering off enough to have things like the XEGS attempting to spur popularity)

And I'm not getting into sprite multiplexing and on-screen colors as that's just going to get into a list of trade-offs and different circumstances. (the A8 can obviously show all 128 colors on-screen and multiplex many sprites as well, but it's more a matter of practical utility and in the case of the A8 the lesser popularity meant far fewer examples and less extreme examples of developers pushing the capabilities -be it graphics or sound)

Also Jack rescuing Atari from bankruptcy/shut down by Warner cost him most of his own personal money, did you expect them to design 16 bit custom chips for $10 in 6 months? Every day he owned Atari he was paying off debts racked up thanks to the rubbish sales of 5200 and A8 worldwide after Warner ran it into the ground, he made it to 1995 from nothing, second time in his business life.

He didn't rescue Atari though, he just found what he was looking for when Warner was offering it. (what he managed was impressive, but saying Morgan would have driven Atari to bankruptcy is hardly accurate -and it seems likely they could have recovered faster under Morgan and been more successful in many respects, so long as Warner didn't screw with things too badly... a shame Morgan wasn't in a position to take on Atari himdelf rather than Tramiel -then again Warner didn't even bother consulting him at all let alone Atari personnel -many of whom sued Atari Inc over the debacle -ie the vestigial corporate shell kept by Warner for legal reasons)

A new computer was needed FAST, GEM is a genius OS superior to Mac, the ST was technically far superior to the Mac but cost 65% LESS. He did a great job in 1985 with the money he had left over rescuing Atari from the creditors. STE was a waste of time though too little too soon. Alienated the public AND the games developers and split his market and increased production costs making STE AND STFM together in factory.

GEM isn't an OS, it's a GUI front-end to CP/M, TOS is a direct CP/M derivative -improved over the older CP/M-68k. (and it's a good thing Atari somehow avoided getting GEM castrated like DR was forced to do with the PC version or the ST would have been much worse off)

AMY was nice but probably cost more than Amiga Paula chip. Antic/GTIA were old news, no place for mono sprites and 160x200 4 colour graphics on 16bit £799 computers in 1985/86

I never suggested using ANTIC or GTIA in the new machine (GTIA would have needed substantial enhancement to really be viable, like double clock modes with double the dot clock for every existing mode at the very least -ie 640 wide 2 color, 320 4 color, 160 16/9 color preferably expanded to 16 color registers at least and either more sprite capabilities or added blitter functionality), but again I wasn't suggesting that. (and Atari Inc already had far, far better hardware to work with than anything on the market) I was saying that with POKEY and the 6502s already in supply (and owning the IP for POKEY) they might have found a useful place into the ST for audio/coprocessing and I/O capabilities (likely clocked at 2 MHz given the master clock in the ST), but it would depend on the cost-effectiveness of their use and practicality of interfacing them with the 16-bit portions of the system (namely glue logic due to the dissimilar architectures) and if it worked you could eliminate the Hitachi keyboard scanner and YM2149 (you'd need a simple 8-bit PI/O port from something else though) and you could use the SIO interface as well (I don't think that could be directly implemented for use in leu RS-232 though) and the timer capabilities. A lot of stuff you could do with a 2 MHz 6502 though, but it would depend on the configuration... preferably at least a small amount of work RAM especially with 2 MHz being too slow to efficiently interleave accesses with the 68k (without wait states) though you might manage 1 access between 68k accesses with how efficient the 6502 is (less than one cycle per access while the 68k only needs the data on the bus 3 cycles after the start of the access iirc and another cycle to complete it), but you'd want at least a small chunk (perhaps 8k) for the 6502 to work in off the bus... (though if interleaving worked it should mean full speed parallel 6502/68k processing, but that would change with other DMA devices on the bus and with faster 68ks unless the 6502 was up-clocked to match) It might even be faster for some graphics related operations even (not sure), but if nothing else it could make for a pretty decent sample based sound system with POKEY's 4 4-bit DACs (or combo sample and chip sounds depending on the case) be it using software timed loops or the 3 interval timers in POKEY (and very efficient 6502 interrupt handling) you've got nice possibilities. (not Amiga quality, but still pretty nice, pretty reasonably 4 8 kHz channels -the lower the sample rate the less the disadvantage of 4-bit PCM to the point of 4 kHz being barely distinguishable and 4-bit samples take 1/2 the space at a given rate) 6502 reads sample from main memory resamples it, plays it via interrupt, etc. (or not resample it and simply use varied playback rate, but while simple, that would eventually get too resource intensive at high playback rates)

A DMA sound chip would obviously be more cost effective (especially one also integrating the YM's I/O tasks ) but they obviously didn't have time to do that and my suggestion is more of a hack of components they had on-hand. (and if POKEY's other features weren't useful or glue logic slowing down development more, it wouldn't have been worth it) Cheaper than many off-the shelf solutions though. (Yamaha's PCM/ADPCM solutions were not even close to cheap)

On sound specifically, most early ST games had sampled sound, Guantlet 1...Technocop....Dungeon Master etc. So whilst the Yamaha chip in the ST was worse than ADLIB in some ways....if you had ADLIB your sound effects were all coming from PC speaker...barely an improvement on a 1982 ZX Spectrum's sound. ADLIB is FM sound I believe, like the SEGA Megadrive, it's not really something you can compare. What I do know is, the ST was more flexible but as time went on and software houses got more greedy less and less sampled sounds were in games and things got worse. You can't play Amiga MODs on a PC without a mid 90s wavetable card like GRAVIS or AWE32 etc, you can on an ST though...bit scratchy sure but you can do it Trouble was in the 80s you need ADLIB for music and Soundblaster for effects. Expensive...2 sound cards for 1 machine. And both were inferior to Amiga and pretty much the C64 SID too. There was talk of getting SID onto an ISA card to sell to PC users as late as 1986.

There were 4-8-channel MOD players in early 90s PC games, and in-game too, not just in the demo (you generally needed a 386DX or higher though) and it's way better than the ST's YM playback with a pure 8-bit linear DAC or up to 44.1 kHz 16 bit with the SB16 and stereo. (with 8-bit it's a bit scratchy as you're adding 8-bit samples to an 8-bit sum while with 16-bit you can add many 8-bit channels and still have plenty of bits left). You didn't need sample based audio cards specifically, just use the SB like the STE/Amiga audio (more or less) though such games often supported tracker/MOD via GUS as well (no need to specifically support AWE-32 or such as those are all backwards compatible). Hell, modern consoles and PCs pretty much gave up on fancy audio hardware (for onboard audio) and opted for plain DMA sound (usually 2/4/6/8 DMA channels often catering primarily to 44 and/or 48 kHz) and from win9x onward MS has supplied a software general midi driver using the Roland Soundcanvas sample set and meeting GM standards, though I'm not sure there were driver to use the SB16's DMA sound for that (quite possible though) and I do know there were drivers to map general midi to the OPL2/3 synth chips (with widely varying results).

If you want examples, try Robocod, Jazz Jackrabbit, Aladdin, or Quik the Thunder Rabbit in for DOS.

Poor analog circuitry and hefty low pass filtering trying to mask that really hurt the quality of many SB compatibles (both FM and PCM) though with a few exceptions like the Yamaha produced cards among others. (emulation generally managed extremely crisp/clean sounding FM and PCM -unfiltered so also a bit scratchy sounding in many cases, but clean- that goes for VirtualPC, VDMS, and DOSBox -they're pretty accrurate too aside from omitting some of the rarely used advanced OPL3 features)

Edited November 1, 2010 by kool kitty89

Crazyace · November 1, 2010

As far as I'm concerned the only mistake or bad decision was not including scrolling in the ST design. ( Having a packed pixel mode rather than bitplanes like the Apple 2GS would have been nice as well ) No scrolling meant that the CPU had to waste a lot of time moving the background - in many cases a 50fps game on the Amiga would be at least twice as slow on an ST - as almost one frame of time would be needed to repaint the background to simulate the scroll.

kool kitty89 · November 1, 2010

The Tramiels were absolutely horrible at any matters concerning hype. Steve Jobs is so good at hype that he is said to generate a "Reality Distortion Field". No amount of technical excellence would have killed Apple's hype. That is just bringing a knife to a gunfight. Without changing a thing, both the ST and Amiga were VASTLY superior in technical terms to the original Macs. Greater technical refinement or better tweaking of the price/performance compromise were not was needed to compete with Job's Apple. What was needed was some P.T. Barnum inspired marketing and a bit more software elegance. Jobs comes at things from a consumer design perspective. He isn't and never was about the feature bullet list.

Except they had limited funds to even create hype and in the extreme case (with the less capable Sam Tramiel handling things) the Jaguar managed enough hype in 1993 to scare major video game companies into thinking Atari was a real threat (of course they couldn't see Atari's fiscal reports ) and got a significant boost of investor interest including a favorable manufacturing deal with IBM among others (the primary reason for pushing the Jaguar in the early test market due to a severe shortage of funds and mounting debt).

Seeing the caused trouble, he better would have Atari let die.

He didn't do much harm, but Warner did... both with A Inc and in the split in '84... (they also drove it to its greatest success in the early 80s, but management problems grew until they became a serious hindrance by '82 and became more obvious to the public in '83 -a real shame Morgan hadn't come in by Summer of 1982).

OTOH Sam Tramiel seems to have had a hand in Atari Corps failngs from 1989 onward. (Jack retired)

In case you haven't been keeping up, Jack didn't "kill" Atari's games and the 7800 likely would have been released in '84 (or at least mid '85) had it not been for Warner contending ownership.

Not all is as it looks for real. Commodore sold the AMIGA just in order of the success for selling the C64. Many people may have thought, the AMIGA is the successor to the C64. At Atari the AMIGA never would have reached this high count of selling.

Yes, perhaps (it's hard to tell what Morgan's reformed Atari might have done), but the real question would be: how would jack have managed the Amiga at CBM?

He may have turned down buying A.Inc on his own in '84 with TTL, but that was a totally different context: he couldn't go buying A.Inc in his position, he had no brand name, no distribution network, and notable (but limited) private funds. Had he been at CBM there

If Atari Inc had stayed (not split by warner) and Amiga not defaulted and lied about their chipset in September, Atari Inc would have had an Amiga based game console out by Christmas of 1984! (the prototypes PCB was all ready to drop in the first run of LSI chips, but Amiga broke the contract and Warner sold Atari Inc consumer just a day or 2 later) But I'm repeating myself from above and Curt and Marty have already established this...

See: http://www.atariage.com/forums/topic/166993-atari-7800-forum-description/page__st__75__p__2123084#entry2123084

http://www.atariage.com/forums/topic/152112-sneak-peak-amiga-atari-design/

So from that standpoint, the Amiga may have sold FAR more than historically and not only a game console but a console intended to expand to a low-end computer the next year (only 128 kB) and following that Atari would have the rights to push the full computer without other restrictions. (so by 1986 the full computer could be launched and Atari could offer expansion beyond 128k for the game systems, which could have meant a A500-like console by '86 and likely using Atari's Unix BDS OS with Snowcap GUI)

But if Amiga did like they had historically and Atari Inc continued under Morgan, they likely would have sued Amiga/CBM and either upheld the contract (no counter suit like the ST) or settled and opted for their own rainbow chipset among others.

That's why I wrote "16 Bit version" . A simple doubling of the clocking frequency and they would have fit to a PC Monitor.

Yes, but that wouldn't address the sprite weaknesses in general, but a blitter would solve that. (though a 3.58 MHz 6502 would also open up software rendering a lot)

But why bother when Atari already had multiple fully prototyped 16-bit chipsets that kicked the Amiga's ass, let alone the ST.

As far as I'm concerned the only mistake or bad decision was not including scrolling in the ST design. ( Having a packed pixel mode rather than bitplanes like the Apple 2GS would have been nice as well ) No scrolling meant that the CPU had to waste a lot of time moving the background - in many cases a 50fps game on the Amiga would be at least twice as slow on an ST - as almost one frame of time would be needed to repaint the background to simulate the scroll.

Scrolling, simple hardware acceleration (line copy, fill, etc) short of a full blitter would have been nice, but how much would chunky/packed pixels have sped up software rendering? (let alone double the CPU speed)

And for non-game applications, do you think that a 256 color display mode could be more significant than added hardware acceleration. (especially if that meant 8-bit chunky pixels speeding up software blits and focusing on faster CPU speeds) Especially after the fact (without hardware acceleration initially) boosting CPU speed would mean all (non timing sensitive) older programs would be faster as well. Not to mention 3D stuff or software scaling and in the context of porting PC games you'd be dealing with software renderer engines as well, and with VGA games especially you'd have 8-bit chunky pixel frambuffers being focused on. (the Amiga version of Wing commander was pretty limited, but imagine a port to a 16 MHz 68k system with a 320x200 8bpp linear framebuffer -did the Amiga port even use the blitter?)

Of course if they were competing with VGA, that did had hardware scrolling as well, but I'd think in many cases the 8bpp chunky framebuffer and 256 indexed colors (at least 12-bit RGB) would be generally more significant. (again, especially looking at non-game applications as well) If they didn't bump to 12-bit RGB they might as well do away with indexing and go to plain 8-bit direct RGB (maybe a toggle for 3-3-2/3-2-3/2-3-3).

Indeed. If the ST had been branded a Commodore (or anything rather than Atari) it would have sold more. Hell, if any of the Atari computers were branded something else, they would have sold more.

In the US maybe. Not in the UK and Europe, Atari didn't have a bad name here. Look at how well the ST sold on these shores

Yes, but Commodore had a much stronger name due to Atari's poorer competition with the A8 than the VIC and C64 and Atari's name was hardly soiled in the US as some seem to imply, a major reason they managed to sell as many 2600s and 7800s in the late 80s was due to the Atari brand name (and tactful management by Mike Katz with limited funding).

How do you think the Amiga would have sold had Jack been there and managed it like the C64 and ST? (1985 Amiga 500 anyone?)

And being branded Atari wouldn't have been an issue either if the A8 had continued to grow in popularity alongside the C64 from '82 onward, but Atari made the wrong move with the 1200XL... had the 600 and 800XLs been what were released in 1982 instead, they could have made for real competition, and of course had Atari operations not been halted in the latter quarter of 1983.

It is a bit ironic that Kassar's dream of an "appliance computer" is almost exactly what the Macintosh became, or even more like the late 90s iMAC. (the PET did that too... actually so did the ST more or less being pretty closed box) Kassar even wanted to push color coordinated models to appeal to women and have smart peripherals where were fully plug and play. Even more ironic that major complaints about lack of flexible user expandability pointed to the Apple II's internal expansion ports when Apple later turned to the appliance computer philosophy. (aside from the high-end workstation units and such -MAC II, PRO, etc)

Edited November 1, 2010 by kool kitty89

Were the Atari ST's big for gaming or just the 8 bit line?

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Join the conversation

Recently Browsing 0 members