Plans in Motion

In the summer of 1970, Nolan Bushnell, Ted Dabney, and Larry Bryan prepared to change the world.  With a vague plan, virtually no funding, and a partnership that consisted of a name and little else the trio began an ambitious project to turn Spacewar! into a commercial product that could entertain teenagers and adults at fun spots and arcades.  Unsurprisingly, virtually nothing went according to plan.  Nevertheless, after months of prototyping, redesigning, and compromising Bushnell believed he finally had his game by the beginning of 1971 before he realized he had made a crucial calculation error and that a minicomputer would not run his game economically after all.  Bushnell’s dreams could have died right then, but in a eureka moment, the budding entrepreneur realized that he and Dabney — Bryan having been abandoned long since — had built such versatile TTL circuitry that a computer was not really needed at all.  After a quick redesign and a fortuitous encounter with one of the only American arcade game manufacturers outside of Chicago, Bushnell’s game, ultimately dubbed Computer Space, was well on its way to entering production.

Despite being a relatively simple game created by just two people over a period of just a few months, there are few games with as controversial and convoluted a history as Computer Space.  This is because for decades the only account of the game’s creation came from Nolan Bushnell.  When Ted Dabney emerged forty years after the fact to pass along his memories of the game’s creation, a completely different story emerged, one which Bushnell continues to vehemently deny.  As with most controversies involving credit and responsibility, the truth probably lies somewhere in between.  In this post, I hope to reconcile these stories as best as I can under the circumstances.

Converting Spacewar!

The Data General Nova 1200, which may have been the target model for Bushnell’s Spacewar! arcade game

In the summer of 1970, the informal Syzygy partnership began its project to recreate Spacewar! as a coin-operated game, with Bushnell handling the electronic engineering, Dabney focusing on video engineering, and Bryan handling software.  The first important task was choosing which computer to use to play the game.  Bushnell has stated in many interviews that it was a sales flyer for the Data General Nova that first sparked his interest in recreating the game for the arcade, but the trio still surveyed the minicomptuer market carefully to make sure that they settled on the cheapest computer with enough processing power handle the job.  According to Bushnell’s 1976 deposition, sometime around August 1970 the trio acquired a catalog comparing the prices and features of all the minicomputers on the market.  After studying these listings, the trio decided the Nova was indeed the proper computer for the task.

According to Bushnell’s deposition, Bryan began work on the computer program even before the choice of computer was finalized, placing this work in mid-Summer 1970.  This also lines up well with Bryan’s recollection that he did his work about six months before Bushnell approached Nutting Associates to manufacture the game.  According to Goldberg’s interview with Bryan, the programmer spent a couple of weeks tinkering with the program in his spare time before deciding that there was no way they would ever be able to get a Nova computer to play four to six games of Spacewar! at a decent speed.  Bryan delivered these preliminary results to his partners and then passed into history. Based on his deposition, Bushnell appears to have continued working on the minicomputer problem for several more months, so he would have probably brought Bryan back into the loop if they had moved forward with that plan, but it was not to be.  Indeed, Bryan must have continued to be on Bushnell’s mind for some time after the summer, for according to Dabney in his Computer History Museum oral history when it came time to formalize the Syzygy partnership, Bushnell brought up Bryan, and Dabney responded that since he never paid in his $100 (which, of course, Bryan maintains that he was never asked to do), Bryan was out of the group.

What happened next depends on whom you ask.  In Donovan’s Replay, Bushnell described a process of whittling down the computer program by transferring more and more functionality like the background star field and gravity to specialized TTL hardware in an effort to salvage the project.  According to Donovan, this work was unsuccessful, and by Thanksgiving the project was dead.  Dabney, on the other hand, has claimed in interviews that work halted completely when a computer program did not appear feasible and did not start up again until one day, out of the blue, Bushnell asked him if it would be possible to control the movement of the spaceships on the television screen using a pure hardware solution.  Based on Nolan Bushnell’s deposition and the accompanying evidentiary exhibits, it appears that Bushnell’s account to Donovan, while not entirely accurate, is closer to the truth than Dabney’s.  From Dabney’s interviews, however, it appears that Bushnell was doing a lot of brainstorming without Dabney and only periodically keeping him in the loop, so it may be Dabney was simply unaware at the time of the effort being put into the project.

Bushnell’s deposition demonstrates that work continued on a computer-based system through at least January 1971, contradicting his recollections to Donovan that the project was dead by Thanksgiving.  The January date is set by one of the few pieces of documentary evidence from the period: a letter from Nolan Bushnell to Data General salesman Bob Washburn dated January 26, 1971, in which Bushnell states his intent to order several Data General Nova computers.  In his deposition, Bushnell explained that by this point the basic system for the game had been sketched out, and he now needed some actual computers to make sure it worked.  Schematics provided by Bushnell at the deposition also prove that several portions of an interface between the Nova and a television had been designed by this point, demonstrating that work had continued on the project above and beyond what Bryan had contributed to the software.  According to Bushnell’s deposition, the minicomputer project only died sometime between January 26 and a followup letter from Bob Washburn dated February 16, 1971, that was also entered into evidence.  According to Bushnell, he initially believed he had solved the timing problems from Bryan’s initial designs (most likely through moving some functions to specialized hardware as he claimed to Donovan since he was not a programmer and could therefore not alter Bryan’s code too much) and was looking to get some time on a local Nova computer to do a final pass on the system.  At that point, a person at the facility to which Bushnell was attempting to gain access pointed out a flaw in Bushnell’s calculations, thereby illustrating that a computer would not work after all.  At this point, Bushnell had a eureka moment and decided to do the entire game with custom-built hardware.

A New Hardware System

Ted Dabney’s House, where the prototype hardware that ultimately became Computer Space was built

Perhaps the biggest bone of contention between Bushnell and Dabney is the creation of the spot generating and motion circuitry that proved critical to abandoning the minicomputer in favor of a custom hardware solution.  In his oral history and Retro Gaming Roundup interview, Dabney claims Bushnell came to him one day and asked him why adjusting the vertical hold on a television caused the screen to move back and forth and whether they could do something similar to move their spaceships across the screen.  Dabney said this should be possible, so Bushnell asked him to build a system to do so.  Dabney then set up a makeshift lab in his daughter’s bedroom to create the prototype.  When it was finished, Bushnell took the circuit board with him and secured a contract to make a game based around this hardware.

Bushnell has vehemently denied this version of events on several occasions, including an interview with this author.  According to Bushnell, he would have never needed to ask Dabney how a television worked because he had been repairing them for years.  Also, he maintains that all of the digital engineering on the project was done by himself alone, and that Dabney was only involved in analog work like a monitor interface and power supply.  Furthermore, he has stated to numerous authors over the years, including Kent and Donovan, that he was the one who set up a lab in his daughter’s bedroom to make the game.  This dispute between the partners is not easy to unravel.

First of all, there is one area where Dabney appears to be mistaken: Bushnell likely never came to him to build a spot generating and motion system to replace a computer.  The January 26, 1971, letter clearly demonstrates that Bushnell was still working on a computer system after the motion circuitry had been developed, so there was no need at the time to do an end run around the Nova.  In his deposition, Bushnell explains that the motion circuit board was an “exercisor,” that is a system built to simulate the actual computer that would ultimately be used.  According to Bushnell, they built this exercisor so they could make sure their other hardware was working properly.  If they just tested everything directly with the Nova, they would not be able to tell if any problems stemmed from the hardware or the computer software.  According to Bushnell, it was only after he learned in late January or early February 1971 that a Nova would definitely not work that he hit on the idea of using the already existing motion circuitry in its place.

On the other hand, it appears that Bushnell’s long-held contention that the early game hardware was built in his daughter’s bedroom is probably a myth.  In his oral history, Dabney states the hardware was built in his daughter’s bedroom, and interviews conducted with Dabney’s daughter by Marty Goldberg and Curt Vendel for their Atari history book appear to corroborate Dabney’s recollections.  Even Bushnell admitted in a Retro Gamer article on the making of Computer Space that at least some of the work was done at Dabney’s house.  Of course, Bushnell also reconfirmed in that interview that he did work in his daughter’s bedroom as well.  Dabney counters that Bushnell’s wife would have never let him bring so much as a soldering iron into their house.  Admittedly, in this case we only have Dabney’s word versus Bushnell’s recollections, so the only evidence we have to go on is the character of the individuals involved.  In the 1970s, Bushnell often took full credit for creating both Computer Space and Pong, which we now know is not true.  Furthermore, he has always claimed that he first saw Spacewar! at the University of Utah, which is also probably not true.  Dabney, on the other hand, never sought the limelight, never sought to take credit for any accomplishments, and only told his side of the story forty years later after he was tracked down by historians.  While this does not automatically render him free of any agenda, it does mean that he is not likely letting a need for self promotion cloud his recollection of events.  Therefore, I conclude, like Goldberg and Vendel in their book, that the preliminary work was done at Dabney’s house, not Bushnell’s.

This leaves open the question of who should get the credit for creating the motion circuit board.  This is a question that will probably never be answered.  Dabney’s contention that Bushnell had no idea how the vertical control on a television works seems a bit suspect, as he did do TV repair work in high school and college.  Even if he was just “switching tubes around” and did not have the knowledge or skill to do sophisticated repairs, I find it highly unlikely that an individual as insatiably curious as Bushnell with an electrical engineering background and direct experience working with televisions would not have any inkling as to how they worked.  Its also worth noting that in his 1976 deposition, he describes how the motion control hardware works and displays an impressive command of the entire system.  Granted, this deposition was conducted years later, so he had plenty of time to refine his knowledge of the technology, but he did very little engineering work after creating Computer Space, so its hard to imagine that he gained most of this knowledge after the fact.  Therefore, Bushnell’s deposition shows that at the very least he was capable of understanding the engineering behind the hardware even if he did not create it himself.  On the other hand, if Dabney’s house was the primary staging area for building the motion circuitry, it would make sense that Dabney led the effort to design it.  Dabney was also the more experienced engineer when it came to working with video circuitry, having spent most of his career at Ampex doing just that.  Therefore, while I am not convinced Bushnell was as ignorant as Dabney claims, I do think Dabney must have taken the lead in developing the motion control board.  A potential smoking gun appears in Bushnell’s 1976 deposition in the form of a drawing entered into evidence that is dated January 26, 1971, entitled “Position and Line Counter Cosmic Combat” that is marked “drawn by S.F. Dabney.”  If Dabney is the one drawing the schematics portions of the spot generating hardware, he is probably also the one designing it.  Regardless of who built it, however, both engineers agree that the motion control board was Bushnell’s idea.

Finding a Manufacturing Partner

Nutting Associates in Mountain View California, where Computer Space was manufactured

Whether built primarily by Bushnell or Dabney, by early 1971 the partners had completed a motion control circuit board that could hook into a television or monitor and generate a dot on the screen that could be moved around by flicking toggle switches on the hardware.  According to an interview conducted with Bushnell by Kent for The Ultimate History of Video Games, the parts for this system came primarily from Ampex, which allowed employees to take small numbers of cheap components for personal projects, and Marshall Electronics, where Bushnell knew salesmen that were happy to provide free samples in return for a promise of an order if the project ever came to fruition.  According to Kent, a used television purchased at Goodwill completed the system.

By early 1971, Bushnell and Dabney also finally had a real company.  While they had been working under the Syzygy name since at least the summer of 1970, they had never done any work to formalize the relationship other than a vague promise to contribute $100 each that was apparently not even conveyed to one of the partners.  In either December 1970 (according to a prospectus prepared by Atari in 1975) or January 1971 (according to Atari’s FY 1973 financial statement) Syzygy Company was formally organized as a partnership between Nolan Bushnell and Ted Dabney.  The catalyst for this event was most likely the impending order of computers from Data General, as Bushnell would have definitely wanted to present a real business entity to such an important supplier.

In nearly every account of the founding of Syzygy/Atari from Cohen to Kent to Donovan, Bushnell and Dabney are said to have contributed $250 each to start the company, though all of these authors have confused the timeline.  Kent, in one of his more egregious errors, states that Bushnell, Dabney, and Bryan came together to found Atari in June 1972 by each agreeing to chip in $250, conflating the founding of Syzygy in 1970 with the incorporation of Atari two years later.  Cohen meanwhile, correctly states that Bushnell, Dabney, and Bryan were all supposed to chip in $100 to create a partnership, but goes on to say that after Bryan pulled out, the other two upped their contributions to $250.  Like Kent, he places these events at the incorporation of Atari in 1972 rather than in 1970-71.  Even Donovan, generally accurate in his recounting of video game history, places these $250 contributions at the founding of Atari in 1972.  A look at Syzygy’s financial statement for the year ending December 31, 1971 and Syzygy’s balance sheet dated June 30, 1972, shows that none of this is accurate.  First of all, the contributions were definitely made at the start of the Syzygy partnership and not at the incorporation of Atari, because the 1972 balance sheet shows that the company controlled cash in excess of $6,000 at that time (derived from several income sources that will be discussed in a later post).  The 1971 balance sheet further proves that the ownership contributions were made during that fiscal year, though it records $350 from each partner rather than $250.  As discussed previously, this has led Goldberg and Vendel to claim in their history that the initial contributions for the company were $350 each.  Most likely, Bushnell and Dabney contributed $100 soon after the informal partnership came into being in summer 1970, as stated by Dabney and implied by Cohen, and then added another $250 each when they formally organized the company at the end of 1970.  This would explain why most accounts say the initial owner contributions were $250 without contradicting the balance sheet that shows $350 in ownership contributions by the end of 1971.

With a company and a prototype system in hand, Bushnell and Dabney now turned their attention to raising the money needed to make their arcade game a reality.  According to Goldberg and Vendel in Business is Fun, the duo first tried to interest Ampex in the project through their boss, Ed De Benedetti, but he turned them down.  Next, Dabney turned to one of his mentors, a former engineer in the Ampex Military Products Division named Irving Roth, but he turned them down as well.  Unsure where to turn next, Bushnell made a fortuitous connection through his dentist.  According to an interview excerpt in High Score!, Bushnell was discussing his project during a dental appointment and learned that another patient worked at a local coin-op manufacturer called Nutting Associates.  According to both Benj Edwards’s article on the creation of Computer Space and Business is Fun (in both of which it must be noted all dates related to Computer Space are off by one year due to faulty recollections of the timeline by Dabney), this dentist appointment took place in February.  Bushnell contacted Nutting, and two days later presented his pitch to create a video arcade game for the company.  By March, Bushnell had resigned from Ampex and joined Nutting, where he prepared to unleash the first commercial video game on the world.