How Atari’s star engineers designed the Amiga (and Commodore’s, the Atari ST)

One of the memories of my geek childhood is, without a doubt, the frenetic rivalry between Amiga and Atari, —which I saw from the sidelines of my Amstrad CPC—. Both systems, launched very closely in time, were called to replace the previous generation, 8-bit microcomputers, in our homes This left a level playing field between the two contenders, the Atari ST and the Commodore Amiga, who were fierce competitors and whose user bases developed the same loyalties and hatreds – intensely and ridiculously – that we associate with Nintendo vs. . Sega of that time or Playstation vs. Xbox today.and in our hearts— and become the kings of digital entertainment. Its main competitors, Apple’s Macintosh and IBM’s PCs, didn’t play in the same league: Macintosh cost four times the price —and at the time, it had monochrome graphics— and PCs were still light years away technically when it came to multimedia capabilities.

Commodore Amiga 1000
Commodore Amiga 1000 – Wikimedia

This left a level playing field between the two contenders, the Atari ST and the Commodore Amiga, who were fierce competitors and whose user bases developed the same loyalties and hatreds —in intensity and ridiculousness— that we associate with the Nintendo vs. Sega feud of that time or the Playstation vs. Xbox of today.

That is why it was very interesting —and enormously ironic— to learn about the tumultuous design of those two mythical systems and, above all, the origin of their respective design and development teams.

The history of the birth of the Atari ST was quite an affair, like everything that happened at Atari at the end of 1983. After the launch of Atari VCS, one of the main engineers of the system —Jay Miner, creator of the TIA chip, the person in charge of the console’s graphic system— began exploring new technologies to use as core for the next Atari console. His attention soon turned to a new CPU from Motorola, model 68000, released in 1979.

Jay Miner con su perro Mitch
Jay Miner, Amiga’s system architecture designer, and his dog Mitch – Amiga User International

Soon, Miner came to the realization that his next-generation system should be based on this CPU, although he also realized that to get the most out of the new system, he would have to back it up with other co-processors that would take away from it some of the heaviest burdens a video game system faced, usually related to transferring large amounts of memory.

At that time, Atari was under a new Direction: During the design of the Atari VCS, Nolan Bushnell —one of its founders and director— closed a deal whereby Warner Communications, the American media giant, bought Atari and gave it a huge financial muscle for the VCS development. One of the main reasons for the deal was that Bushnell and his team were overwhelmed by competitive pressure —Fairchild had launched the first console with a microprocessor and the result was frankly impressive— and had decided to get financing at any price.

For many years, Atari was one of Warner’s subsidiaries.

However, after 1983’s videogame industry crash, neither Warner nor Atari were happy with the deal. Warner viewed Atari as a bottomless pit bleeding capital at a rate of $10,000 a day, while Atari viewed Warner as an impediment on their attempts to regain leadership in the industry. Faced with the situation, Bushnell left the company in 1979. He would not be the only one: his replacement, Ray Cassar, was soon infamous for his bad treatment towards engineers and developers. Among other things, and to save costs, Atari began to swing the lead on the bonuses promised to engineers. The combination of factors caused the rout, and among those who left was the group known as “The Gang of Four,” who left the company to found Activision, with Larry Kaplan among them.

When it came to the new console, contrary to Miner’s ideas to design a revolutionary new system, Warner had other intentions: to design another system based on components that were as cheap and readily available as possible. This, of course, was far from involving a next-gen CPU, much less the design and manufacture of specific chips. Faced with the situation and the deception with his annual bonus, Miner would leave Atari.

But Miner had it in him, and after a brief period, ended up again founding a game systems development company called Hi-Toro, with Larry Kaplan. The purpose of Hi-Toro was the design of a next generation system, more specifically the one dreamed by Miner, based on the Motorola 68000, whose financing would come from the design and manufacture of components and peripherals for videogames.

Motorola 68000

Once at Hi-Toro, Miner created his basic architecture based on the 68000 and three accompanying chips of his own design: Denise, Agnus and Paula —tasked with taking the heavy weight off of multimedia work, mainly involving large sound-related memory transfers and graphics, to the processor—.

Although Miner had a personal computer in mind, investors viewed ‘Lorraine’ —Miner’s architecture— as a next-generation console: not surprisingly, the console market was extremely lucrative. Miner, who did not want to surrender his ideas, contemplated in his architecture a set of expansion ports so that this console could easily be transformed into a personal computer. This decision would be very successful, since after 1983’s crash, consoles had totally lost their appeal to investors, while personal computers had become the new trendy technology investment.

In its presentation at CES in Chicago in 1984, the new architecture would cause a sensation and attract the interest of Atari, whose plans to bring out another continuity system to the market as cheaply as possible had become obsolete in light of the debacle of ’83. Atari then invested half a million of dollars in Hi-Toro so they can keep working while they close a larger deal to use the Lorraine architecture to develop their next generation system.

Jay Miner’s Lorraine Architecture – Ungzd/ Wikimedia

The timing can’t be better, as Hi-Toro had run out of money, the company was desperate and its members had mortgaged every last of their possessions. Miner even receives visits from Steve Jobs, which did not lead to any kind of collaboration. So Hi-Toro accepts that sum in exchange for giving Atari the use of its designs for a year, while a deal is closed whereby Atari ends up buying the majority of Hi-Toro’s shareholding. That money s a lifeline for the company.

But the truth is that Atari, in reality, wants to take control of Hi-Toro —now renamed Amiga Corporation to avoid conflict with a Japanese brand— to get Miner’s chips: Agnus, Denise and Paula. The half million dollars is a poisoned gift: Atari knows that Amiga will not be able to get the necessary money to repay that loan, so it will end up controlling it.

But in that same year, 1984, Jack Tramiel, CEO and founder of Commodore, had a serious disagreement —the confrontations had been going on for a long time— with the company’s main shareholder, that cost him his position. Tramiel left the company and with a combination of vindictive lust, entrepreneurial spirit and a bulging bank account, bought Atari from Warner Communications, which after the bleeding of money that the subsidiary had caused the previous year had lost its taste for video games.

Jack Tramiel in 1984 –

But Tramiel was in fact only interested in the brand, its prestige and its many contacts and business relationships. “Atari Inc.”, as it was known until that moment, is extinguished and its 10,000 employees end up on the streets. Tramiel then founded “Atari Corporation” with 900 of those former employees. Meanwhile, at Tramiel’s previous company they do not forget Tramiel: To top the removal of their former president, when at Commodore it’s learnt that Tramiel is interested in buying Atari, they approach Amiga and give then a million dollars, more than enough to pay back the Atari loan and get rid of its influence.

Amiga, until that moment Atari’s hope for the short-term future, joins Commodore. When Tramiel discovered what had happened, he sued Amiga and put the Atari’s remnant team to work on the design of a new machine, which must compete with Amiga’s. Among this team were, ironically, Commodore’s engineering heavyweights, who had followed Tramiel on his way out of the firm.

Among others, the Atari team included Shiraz Shivji —one of the engineers who had worked on the popular Commodore 64 and who became the new Vice President of Research and Development— and Tramiel’s own son, Leonard, a talented engineer that had greatly contributed to the VIC-20 at Commodore and became the head of the Atari Operating System team —and who was said to be the only person who could deliver the bad technical news to Tramiel without being fired—.

Shiraz Shivji, one of the fathers of the Commodore 64 and the Atari ST –

For the new architecture, Atari took into consideration a design using 32-bit processors from National Semiconductor, but neither the performance, nor the price, nor the manufacturing capacity of NS convinced Atari, so the final design was based on the one that was would become the most popular chip of the generation, the Motorola 68000. As in the case of Amiga, the processor had several support chips —cryptically called GLU, MMU, DMA and Shifter—, which unloaded it from tasks such as screen synchronization management, memory address range decoding, peripheral management, display RAM or high-speed memory transfers.

Even so, the graphic hard work was still a CPU task, as it did not support sprites or hardware scrolling. Meanwhile, the capabilities of Amiga were far superior to those of the Atari ST thanks to components such as its blitter, DMA or —particularly— the one known as Copper. These allowed Amiga to make large memory transfers in synchrony with the screen’s electron beam, something Atari would try to compensate by adding a blitter to its later STE range.

Atari ST
Atari 1040 ST – Wikimedia

In short, Atari and Amiga were systems with many similarities, but also many differences that in the end marked the future of each of them. While the Atari ST became the computer of musicians —thanks to the fortunate decision of including MIDI ports— and became popular in the world of desktop publishing, the Amiga became the computer for multimedia and video games, the gaming machine that everyone wanted to have at home.

The birth and design of both machines were strongly intertwined with and influenced by the turbulent times in which they happened, to the point that —and that was the point of this too long of a post— the Commodore machine was designed by what had been the engineering team of Atari and Atari’s by the Commodore’s engineering team, in a kind of non-friendly and ironic exchange of talent. Something that was not that extravagant at a time, the first 80s, in which the industry and the technology evolved as much from the power of ego, envy and revenge as from sheer genius and passion.