It is hard for me to be objective about the Atari ST, because I was a dyed in the wool Amiga fanboy in the early ’90s. But the Atari ST was released April 8, 1985 and quickly sold 50,000 units. For a while, it looked like the future belonged to the Motorola 68000-based computers and the ST was going to be a big part of that.
Then again, since you probably are not reading this on an Atari ST, there is an argument that the ST was a failure. So it could be that I am grading on a curve. I expected the machine to be a nothing burger, and it ended up selling 2.1 million units at a time when 2.1 million units shipped still was a pretty impressive number.
I look at the Atari ST as the 1969 Chicago Cubs of retro computers. It didn’t win a championship of any kind, but it hung in there and it’s easy to see how it could have won if a couple of things had gone differently, so people remember it fondly. It was a strong comeback for a legend who had fallen on hard times.
Why the Atari ST was a big deal when it was released
It is pretty easy to see why 2.1 million people thought highly enough of the Atari ST to buy one. It gave a Macintosh-like experience for the price of a PC clone. And it came on the market in mid-1985, barely a year after the Mac, and it beat the Amiga to market by about five months.
It cost 1/3 as much as a Mac, and there is absolutely no way you can argue it was only 1/3 as good. It had the same CPU, it had more memory, it had color, it had a better sound chip, and it had built-in MIDI. And its slick industrial design by Ira Velinsky looked amazing.
The Atari ST was almost everything the standard 1991 386SX PC clone running Windows 3.0 was, but in 1985.
It looked like the future. And unlike the Amiga, which failed largely because it was too far ahead of its time, the Atari ST was a better match for its time without looking stodgy and old-fashioned. The industry looked a lot different all of a sudden in 1985 than it had in 1984, and the ST was a big reason for it.
Like I said before, I’m not an Atari fan. I look at the ST as something that could have done much worse, but in a fair and just world, it deserved to do much better than it did. Jack Tramiel didn’t understand that to conquer the computer market, you had to do a something more than beat everyone on price.
Where it succeeded
Initially the Atari ST sold well. For someone who wanted the ultimate home computer experience in 1985 or 1986, the ST was a good choice. The main thing holding the ST back was that none of the large computer store chains wanted to deal with Atari or Jack Tramiel, so it was a specialty shop item. And with no mass market appeal, attracting software developers was hard. The ST did attract some developers early on, but it couldn’t keep up momentum. The conventional wisdom is software piracy killed the ST, but I really think it was the combination of software piracy, the system not being easy to buy, and Commodore getting its act somewhat together with the Amiga after 1987. Had Commodore and Atari not had to split the PC and Mac markets’ leftover scraps, maybe ther would have been enough there to support a third computer platform.
Thriving in a niche
The Atari ST really thrived as a music workstation. It had built in MIDI, and the MIDI worked really well, so musicians used it. The ST was standard equipment in recording studios into the early 90s. Apple eventually displaced the ST in recording studios, but in the 1980s, there was no reason for a musician to prefer a Mac. The 68000-based machines all had their specialty. For the Mac, it was publishing. The Amiga ruled video production. And the ST ruled music.
The Atari ST’s legacy
Over the years I’ve met several people who owned STs. They love them for the same reason I love the Amiga. They moved on to PCs like I did, but for a long time, there was something missing. That’s because in some ways, the ST and Amiga were better than a PC running Windows 3.x. If you didn’t experience it, that statement sounds absurd. But if you experienced it, you totally know what I mean.
The ST was flawed, but the PC that came after it was too. Both of them were flavor of the week material more so than marriage material. And as a flavor of the week, the ST was more fun in some ways than Windows 3.0 or Windows 3.1. The newer tech was fun too, but there was something about the ST that keeps you wondering about what might have been, if it had been the one that got the chance to mature.
David Farquhar is a computer security professional, entrepreneur, and author. He has written professionally about computers since 1991, so he was writing about retro computers when they were still new. He has been working in IT professionally since 1994 and has specialized in vulnerability management since 2013. He holds Security+ and CISSP certifications. Today he blogs five times a week, mostly about retro computers and retro gaming covering the time period from 1975 to 2000.
Exactly: ST was what 386 with Windows 3 will be in 90s (although regarding MIDI it will take another decade for PC to become good replacement for ST).
Compering to Mac, ST also had one big advatage: it was 1/3 faster than Mac for 1/3 of the price!
Original Mac had to stall CPU 68000 when graphics was drawn on display since memory speed was not sufficient. ST did not suffer from this problem!
Also, ST had ultra fast DMA port, later Atari release ultra cheap laser printer that use ST CPU (similar to modern printers) and lack it own memory and CPU so it was few times cheaper than Apple. This was very popular combination for DTP.
I did not realize the ST also had such a big speed advantage over the Mac. I know the Amiga did too but I attributed it to the coprocessors. The Amiga also didn’t have to stall its CPU when drawing the screen and that probably made a bigger difference.
Atari ST actually sold 3 to 4 million units according to ST Format magazine at the time 🙂 It was a great system which I used a lot to earn money by converting hand written notes from a local musician into printed sheet music with Notator 3 SL. Did also some DTP stuff with it. System had a very sharp high res monitor that was easy for eyes.
The 2.1 million figure came from Jeremy Reimer’s Ars Technica article on computer sales figures, at https://arstechnica.com/features/2005/12/total-share/
I completely agree, the ST was a great system and very underrated.
I too was an Amiga fan, moving from an 8-bit Radio Shack Color Computer to the Amiga 1000 in late 85. The price of the Amiga was significantly beyond my early 20s budget so I was only able to finagle one by finding someone who needed some software written on the Amiga and getting them to buy me one in exchange for writing what they needed. But such was the siren song appeal of Amiga’s promise in mid-85. I spent months poring over every page, image and word in the “Launch” issue of AmigaWorld Magazine, which was in reality a cleverly disguised extended sales brochure that came out months before the computer itself was available for retail purchase.
No teenager ever inhaled every inch of a Penthouse magazine in the detail I memorized that issue of AmigaWorld. It was truly computer porn in every sense, an airbrushed fantasy which significantly surpassed the reality of the computer that actually shipped for at least its year on the market. There wasn’t much you could do with a $2000 Amiga 1000 system in the first months other than run Boing, RoboCity and other demos ($2000 including the “optional” chip RAM upgrade (which was in reality required), monitor and external 2nd floppy drive). The early launch applications like Graphicraft weren’t quite complete enough to be useful for much real production work, largely because when Amiga 1000 shipped the paint was still wet on the operating system itself. Worse, Addison Wesley the publisher of the official developer docs for the Amiga took their sweet time actually printing and shipping the books, despite the fact it was only a re-layout of the docs Amiga supplied in Xeroxed form to early developers. Unable to wait any longer as my Amiga-purchasing benefactor needed their software, I drove three hours away late at night to the house of a developer that had the original Amiga docs and took them to an all night Kinko’s and spent the hours between midnight and dawn copying every page by hand. But… strangely, we loved the machine anyway. Thankfully, a year or so later the OS had matured enough and real tooling, apps, docs and source code examples (in the form of Fish Disks) started appearing enough that the Amiga’s fantasy potential slowly started to become real.
You’re right that most Amiga-centric people from back in the day may simply be incapable of assessing the Atari ST in a completely fair and balanced way. It may be because we didn’t assess the early Amiga in a fair and balanced way either (but in the other direction). So, congrats on setting aside your own ingrained worldview to put the ST in it’s rightful historical place.
What a fun Amiga story! Of course I’ll have Amiga content later in the year… I spotted an ST in a used computer store years ago, with no price on it. Now I wish I’d asked.
My friend Karl Kuras did a superb interview (link below) a couple of years ago with Leonard Tramiel, Jack’s son and the head of Atari’s software group during the 520ST days. One notable thing Leonard said was that Atari initially approached Microsoft about adopting Windows to the ST but that they wanted something like a year of lead-time to build it properly. Jack wanted an OS within a couple of months, which is why they ended up signing a deal with Digital Research to have them create a hasty port of GEM.
https://videogamenewsroomtimemachine.libsyn.com/leonard-tramiel-part-2-atari
Jack couldn’t get out of his own way when it came to screwing over software developers. One story mentioned on “Computer Chronicles” was that Tramiel actually charged 45 developers $1,000 in “rent” to feature their software at Atari’s COMDEX booth. Atari was the only company that essentially made a profit on its own booth.
Along similar lines, Jack appeared before a Software Publishers Association meeting before the ST’s launch, begging them to develop for his system. Except those that agreed soon found they were charged $4,500 for a development system. As a result, there were few companies willing to develop before the machine launched. (To be fair, Apple pulled similar nonsense with the original Mac launch.)
I’ve heard the story of how the ST almost ran Windows, but DRI could meet Atari’s deadline and Microsoft couldn’t. I remember reading it in print years ago but have no idea where. Good to hear Leonard Tramiel confirming it, now I can just quote him. And I also heard the stories about Jack Tramiel profiting off the software developers. Short term gain, long term loss.
No doubt Jack Tramiel was a nnotorious cheapskate and not unaccustomed to screwing over partners but, as someone who exhibited at Comdex many times over the years in increasingly larger booths, Atari almost certainly didn’t *make* money from that deal.For more than a decade, Comdex was the largest, most expensive trade show in the world. Some years my company’s Comdex booth cost over $300,000.
Also, it wasn’t unusual for companies to ask third-parties to help offset the costs of the booth in exchange for some space. $1,000 to get a spot big enough for a demo computer and monitor along with a small sign with your company name in a brand-name, high-traffic booth everyone could easily find would have been a pretty good deal. You could book meetings with potential partners, resellers, press people and key customers. During the peak years, Comdex exhibit space was completely sold out to regular, long-time exhibitors. New, smaller companies couldn’t get any space for love or money and were relegated to trying to convince important people to get in a taxi and travel to a distant hotel to visit you in your suite. That usually resulted in nearly 100% no-shows for meetings (even getting cab in Vegas during Comdex could be impossible).
Last, it was quite common for manufacturers to charge for developer kits. Very few devs got free dev kits, and the ones that did had serious leverage, best-selling IP and a huge international brand. I’ll certainly agree that most dev kits were really expensive, but from the manufacturer’s perspective I don’t doubt they still lost money on every one, even at those prices. Those dev kits were small volume, early units that were hand-assembled and tested – often by senior design engineers because there was no manufacturing line yet. I worked at by far the single most important third-party developer for a large computer company’s platform and when they made a new generation, they gave us ONE dev kit for free (because our products sold literally millions of dollars of their computers for them). We had to buy the other five dev kits we needed.
how does the motorola 68000 instruction set vs x86 instruction set
how does the Atari ST compare with the Color Computer 3
instead of the Tandy 1000 what If Radio Shack offer a “Color Computer ST ” with Atari ST motorola 68000 spec/clones at same price point to Atari ST
The CoCo 3 and Atari ST aren’t really a fair comparison, the ST had a faster processor, much higher max memory capacity, better sound and more colors. You’re comparing across generations and the newer generation is going to win. Tandy would have had trouble cloning the ST’s video chip because it wasn’t an off the shelf part. Tandy was right to release the 1000, they just waited too long to switch to a more conventional PC. From about 1985 to 1988 or so, the Tandy 1000 outsold everyone else.
As for the 68K instruction set vs x86, I’ve done very little assembly programming on x86 (literally one very simple program) and none at all on 68K so I don’t think I can do a good compare/contrast.
Back in the day, I owned a Tandy 1000, but I visited my local Radio Shack in a mall, where they also had the Color Computer 3, about the same price as Commodore 64 at Toys R Us.
iirc Color Computer 3 was also sold around 1985 at about $200 but expanding it would be close to the ST in price
Tandy/Radio shack back in the day, could have, instead of introducing the Tandy 1000 and MS-DOS, or even Color Computer 3, created a 68000 computer with Atari ST like specs and video and sound at ST price point iirc around $799.
re: Tandy would have had trouble cloning the ST’s video chip because it wasn’t an off the shelf part.
Couldn’t Tandy use an off the shelf part, similar to perhaps the Color Computer 3 video chip, it wouldn’t be software compatible with the Atari ST, but Tandy could license or create their own operating system for their system. Something like porting OS/9 from microwave to their hypothetical 68k with GEM desktop, ideally with a GPU for sprites and scrolling.
wasn’t OS/9 extremely advance for its time? better than Amiga OS, ST TOS, and even Mac OS and DOS.
it wouldn’t be exactly like the Atari ST, only similar specs like 68k cpu and 1mb of ram at a $799 price point, using only off the shelf components to cut costs and R&D. It could be somewhere between Amiga and Atari ST in features performance and price, but sold by Radio Shack. it would be another 68k machine alongside Apple Mac, Amiga and ST and even Nextstep.
Tandy Advance Computer 68k home computer, by not selling the Tandy 1000 or Color Computer, this 68k based Computer I think could have sold well if it had better graphics and sound.
Tandy also offered the Tandy 3000hl and 3000 series with real 286 and AT architecture for businesses.
I had and used several Color Computers from 1981 to 1985 (when I moved on to the Amiga). Today I still own at least one of all the various versions of the Color Computer, including the Coco 3. Unfortunately, the Coco 3 was too little and shipped too late. Based on the then 8 year-old Motorola 6809 CPU, it was an 8-bit computer shipped when the world had already moved on to 16-bits in PCs (8086), Mac (68000), Amiga (68000) and Atari ST (68000).
Launched in only a 4K RAM version in 1980, the Coco line became quite successful for Radio Shack but to understand Radio Shack computers it’s important to know that the parent company Tandy wasn’t ever really a computer company, they were a retailer who mostly got manufacturers to make stuff under the Tandy brand that they’d then exclusively distribute in Radio Shack. Tandy never really built up an internal competency around computer design and they mostly outsourced computer manufacturing to third-parties (with a few notable exceptions for limited periods). Tandy didn’t even design the original Coco. It was almost exactly a reference design created by Motorola to show potential chip buyers an example of a computer that could be made entirely with Motorola’s new line of 68xx chips (primarily the 6809 CPU, 6883 memory controller, 6847 video controller and 6821 peripheral controller). Tandy didn’t write any software for the Coco either. Instead, they hired Microsoft to port their ROM BASIC to the Color Computer (which was the last ROM BASIC Bill Gates personal wrote some code for). There was literally nothing exclusive to Tandy in the Coco other than the plastic case and manuals. In fact, the unrelated UK company Dragon Computers used the same Motorola reference design and a very similar Microsoft-licensed ROM BASIC to make their own 6809 computer in the UK at about the same time. This wasn’t a Coco clone, it was two companies assembling a computer from available third-party designs and components. And the Dragon ended up being largely Coco compatible as a result (I have a couple of Dragons too).
So, at Tandy the Coco effort was quite unlike Commodore or Atari who staffed large permanent teams of computer and chip designers and rolled their own exclusive custom chips for graphics and sound. Unfortunately, this continued on as the Coco evolved into the Coco 2 which was the same hardware design in more modern plastic with a better keyboard and more RAM. Then they played around with various concepts for a somewhat upgraded “Deluxe Color Computer” but it was never green lit or announced because they couldn’t figure out a cheap, easy way to give it competitive custom graphics and sound hardware, since they had no internal chip design expertise of their own. Eventually, they killed that idea and started on the Coco 3 design. Unlike the Coco 1 and 2, the 3 does have some custom hardware but it’s minimal and far from state of the art for late 1985. They hired an external contractor to roll the graphics in a custom gate array but the design had none of the cleverness of the Amiga custom chip architecture like blitter, copper and HAM graphics mode. So, even at launch, the Coco 3 didn’t quite manage to even reach the level of the graphics and sound hardware in the then four year-old C64 or Atari 800. It did have a couple higher resolution text modes but there was nothing special or uniquely powerful about those modes either.
The Coco platform was great because it was inexpensive, available in 7,000 local stores, had a very good ROM BASIC, *great* beginner manuals and the fastest, best 8-bit processor ever made. The 6809 CPU was substantially better than the 6502 or Z80 and is, even today, widely considered the ultimate 8-bit CPU. Its addressing modes and orthogonal instruction set were more advanced than any other 8-bit CPU because it was an 8-bit/16-bit hybrid and was literally the “little brother” of the much-beloved 68000 CPU (a 16-bit/32-bit hybrid). Apple’s initial Macintosh designs were even based on the 6809 before they migrated to the 68000. Unfortunately, the Coco 1 and 2 had no custom graphics or sound hardware because Motorola never made anything other than the most basic bitmapped video controller chip (with no sprites and no cool mode tricks) and Radio Shack didn’t have the ability to design their own graphics chip or sound synthesis chip like the C64’s legendary SID chip.
very interesting. i saw the CoCo3 in Radio Shack back in the day. How does CoCo3 with 512k compare with C64?
I saw a book “CoCo: The Colorful History of Tandy’s Underdog Computer”
couldn’t Tandy use Motorola reference system for 68k for their
“Deluxe Color Computer” so it would have specs similar to the Atari ST, but running OS9
That’s a good question. I’m not aware of any Motorola reference design for the 68K but I haven’t ever looked for one either. Such reference designs aren’t usually public, they’re used as support materials in engineering level sales calls with companies who might make products using Motorola products.
Circa 1985 Radio Shack’s biggest computer revenue was selling their Tandy flavor of PC compatibles to small business. These machines were generally quite competitive on price and features and were where the vast majority of Tandy’s design effort and expertise went. Around that time any 68K computer would have to sell for around $1000 MSRP. The 1040ST was considered amazing for offering a 16-bit computer with a megabyte for only $1 per K of RAM. The original Mac was launched at $2500, so the 68000 CPU was expensive, flagship stuff then.
The problem is to Tandy the Coco products had always been their cheapest real computer and targeted strictly at home consumer users (not counting the Pocket Computers which were souped up programmable calculators private labeled from Sharp and Casio). The Coco marketing materials only had photos of kids playing games and mom in the kitchen looking at recipes loaded from cassette tape (seriously). The 128K Coco 3 launched in mid-86 at a price of $220. Matching a 68000 machine with less than 512K would be a weird mismatch (even Atari and Commodore bailed on 256K versions of the ST and Amiga). And it would be impossible to do even a bare bones 68000 with 512K for less than $700 or $800. Even the 520ST launched at $800 and Jack certainly cut every dollar he could (and, too often, more than he should). So a 68K Coco 3 would have been non-starter because over $500 it starts to compete with Tandy’s main line PC-compatible computers. Frankly, the vast majority of salespeople and execs in Tandy’s computer business mostly ignored the Coco. If you went into one of Radio Shack’s dedicated Computer Centers, which only sold computers and were considered computer flagships, you were lucky if you even found a Coco set up on demo in the back.
As for comparing the Coco3 with the C64, I don’t feel qualified to weigh in since I never used either one back in the day. Although I now have all of the various models of both lines in my retro collection, I didn’t acquire those until the late-90s.
The ST always left a bad taste in my mouth because it was not backward compatible with my beloved 8-bit Ataris (I have an Atari 800, still hooked up, and it works great)
But my brother bought an ST in 1989 (the shop owner showed him porn pictures to demonstrate the graphics quality) and it had a lot of great games.
A few years back, I got a working ST with a ton of software—not just games but everything, including MIDI support. At some point, I will unlock the computer’s full potential…not just play Colonial Conquest.
A forgotten aspect of computer history is that the ST was actually the leading 16-bit platform in the UK (and probably several other European countries) until 1989.
Until David Pleasance’s Amiga sales packs launched, the ST was significantly cheaper with significantly more software. It was a no-brainer, especially for budget conscious buyers.
Not to mention the ST’s hi-res mono graphics mode, which was great for productivity, and of course the midi ports that established the ST as the musician’s computer of choice.
1989 was definitely a changing-of-the-guard though. “Shadow of the Beast” showed off the Amiga’s true potential. The Amiga Batman Pack was marketing genius. And Commodore cut its prices. Atari responded with the STE but it was too little too late. The final iteration of the ST, the Atari Falcon, was actually great but by late 92 the market was beginning to lose interest in wedge-shaped home computers.