WHILE THE SOVIET UNION BASKED IN THE SUCCESS OF SPUTNIK and continued test-launching its intercontinental ballistic missiles, it was still significantly uncompetitive in the field of computing.
The buzz of competing ideas in both universities and the private sector that characterized the American effort to build bigger and better computers did not exist on the other side of the iron curtain. The Soviets mostly built machines to solve specific problems. They had one that was attempting to predict the weather, and another that was dedicated to attempting to crack the problems of machine translation.
The Soviets had long suspected that America had the edge on electronics. They had seen how the F-86 fighter had outmaneuvered the MiG-15 in the Korean War. Reluctant to believe that American engineering could be superior, the Russians had concluded that its superiority must come from its significantly more advanced on-board instrumentation.
The upper echelons of the Soviet science community knew about America’s digital computers. The general commentary on computing advances in newspapers and magazines would have been sufficient to comprehend the technology gap that had evolved.
They wanted to know more. Since Khrushchev had come to power, bringing an “intellectual thaw” that had run through the Soviet science community, academics were permitted to have greater access to research and ideas generated outside the Soviet Union and its Eastern bloc neighbors.
In the autumn of 1957 a group of Soviet computer scientists applied for visas to travel to the United States in December of that year. They asked to attend the Eastern Computer Conference in Washington, DC—one of the prime forums for discussing cutting-edge computer technology. Surprisingly, the US Department of State thought it would be a good idea to allow them to attend, on the condition that a corresponding exchange be arranged with the Soviet Union. In anticipation of a trip to Moscow, a small group of American computer scientists was assembled by the Institute of Radio Engineers and warned that they should be on alert to travel at extremely short notice. But no invitation came from the USSR in time for the exchange to be arranged ahead of the conference, and the whole thing was dropped.
Nonetheless, a few days after the conference, the National Joint Computer Committee—a body formed from representatives of the trade organizations for electrical engineers and radio experts and the recently created Association for Computer Machinery—held a meeting to discuss an exchange program. They voted in favor of inviting a Soviet delegation to attend the same conference the following year, in December 1958, when it was due to be held in Philadelphia. In addition, they proposed to take the Russians on a tour of American computer factories and research labs. Naturally, their thinking was that the Russians would extend the same courtesy to American scientists. It was an offer of information exchange, similar in concept to the Open Skies proposal that Eisenhower had put forward two years earlier.
Mort Astrahan, a top executive at IBM who chaired the National Joint Computer Committee, was tasked with leading the negotiations. Astrahan was intimately involved in the development of the SAGE air defense system and subsequently worked on the first IBM computers controlled by a typewriter-style device. To discuss the idea of a Soviet information exchange, Astrahan contacted the State Department for help, where he was put in touch with the East–West Contacts staff, who started to advise on the logistics of such a trip.
Astrahan was told that it would be up to the committee, as hosts of the Russians, to make travel arrangements, find guides and interpreters, and get clearances from all the installations that would be visited. The committee would also be responsible for ensuring the delegation did not stray into any closed areas.
The State Department also furnished Astrahan with a long list of all the things the committee should ask to see in the USSR in return. Armed with this information, the committee drafted a letter to the Soviet Academy of Sciences in April 1958 suggesting the exchange. They received no response.
In the meantime, an academic at the University of Michigan successfully hosted a group of four Soviet academics for four days at its annual conference on digital computers in June 1958. The trip was directly reciprocated, with four academics from the university making a four-day trip to the USSR in August; they visited a number of the Soviet computer factories and research institutes and lectured on the work being done in America. Although they were on opposite sides of the Cold War, the trips were conducted in the spirit of academic cooperation.
While in the Soviet Union, John Carr, the leader of the trip, broached the subject of a bigger exchange with Sergey Lebedev, the vice president of the Soviet Academy of Sciences and the nation’s top computer expert. Having already seen the benefits of an exchange, Lebedev was keen on the idea. The invitation was re-sent, but this time cabled directly to Lebedev, and followed up with a registered letter.
By October, Astrahan received a response—though he had some difficulty understanding it. Not only was it in Russian, but the Russian characters used to write the message had been garbled in the process of sending the cable. Astrahan needed help from the State Department just to get a sensible translation. The message was a long list of demands for access to American facilities, particularly the IBM installations working on the 704 machine and the newer 709. They also wanted the full program of the Eastern Computer Conference. There was no mention of a return invitation.
Given that it was already October by the time the Russians responded, organizing anything in time for the December conference was going to prove impossible. The committee decided instead to arrange a tailor-made trip for the following spring, arranging visits to almost all of the facilities the Russians had asked to see. After fourteen telegrams and four more formal letters, the trip was confirmed and its itinerary agreed upon.
Included on the itinerary was a trip to Boston to see, among others, the young assistant professor at MIT whom the Russians believed was building the guidance system for America’s intercontinental ballistic missile. They had seen from the documents they had been sent that Dudley Buck was one of the expected headline speakers lined up for the conference in Philadelphia.
The conference went forward without the exchange, and Buck was indeed a star performer. He gave several talks during the event, which ran from December 3 to December 5, 1958, covering his work with self-organizing systems and trends across the nascent computer industry.
Yet the main paper he presented was the one that really left a mark. Titled “An Approach to Microminiature Printed Systems” the paper, authored jointly with Ken Shoulders, explained in detail how to make a microchip. It explained the work he and Shoulders had been doing in the lab with the electron gun to make the new, ultrafast, thin-film cryotrons.
Shoulders had just been lured to a new position at Stanford University to work in its famous research institute. He and Buck continued to work together on developing the cryotron, but it became a long-distance relationship.
Stanford was extremely keen to bring Buck to California, too. John G. Linvill wanted him to take over as the university’s head of electrical engineering. It would have been quite the promotion. Buck had turned down offers of staff jobs from the University of Washington and the University of California, but Linvill was offering the chance to run an entire department and to be reunited with Shoulders.
Buck took Jackie with him to visit Linvill. They even picked out where they would like to build a house—the job came with a free lot in the middle of what is now Silicon Valley. Yet the new academic term was already under way by the time Stanford made its approach, so Buck took the view that there was no need to leap to an immediate decision on whether to accept.
The Stanford job was appealing. At first glance, it seemed a more supportive place. Shoulders published a paper through the Stanford Research Institute in October 1958, not long after he got there, outlining the work they had been doing with electron beams and thin films of metals. The day after it was published, Buck gave a talk to assorted senior professors at MIT, and the patent experts from Research Corporation, the agents that commercialized the institution’s patents. They viewed it as a “blue sky” presentation, and seem to have been unsure as to what Buck had actually achieved.
Perhaps it was indicative of the cultural differences between America’s conservative east coast and the unbridled optimism that characterizes California in particular. Yet it appears that there was an intellectual elite within MIT that was never quite willing to accept Buck’s achievements.
There seems to have been a view at MIT that the hype surrounding the cryotron was overblown. Perhaps the litany of highly-decorated professors with their well-documented wartime achievements were reluctant to have their place in the spotlight taken by a young experimentalist who was still lacking in formal academic achievements.
Even Gordon Brown, the outspoken Australian who headed the electrical engineering department, and who had previously agitated for Buck’s work to be properly recognized, appears to have been of this view. The Franklin Institute in Philadelphia wrote to MIT in November 1957 asking for a specimen example of the gadget to form the core of a planned cryotron exhibit. Brown eventually turned down the request after ignoring the museum’s correspondence for as long as possible.
Thanks to his prizes and media attention, however, the outside world continued to see Buck’s cryotron as a seminal invention. McGraw-Hill, the educational publisher, sent Buck a check for thirty-five dollars in July 1958 asking him to write about it for their latest encyclopedia of science and technology.
By August 1958 the US Department of Defense had taken delivery of the prototype cryotron computer that had been built for them by Arthur D. Little. So far as Buck was aware, it was not yet being used for any specific military purpose. Yet it contained “1,800 cryotrons, which, themselves, occupy 10 cubic inches,” wrote Buck in response to one of his students. “The control equipment and Dewar vessel occupy 70 cubic feet.”
The Arthur D. Little computer had been built using the wirewound cryotrons. Buck, Chuck Crawford, and Ken Shoulders were rigidly focused on perfecting the new iteration of the design, produced through electron beams. Thanks to the powerful microscope they were using, and the tiny electron beams, they could see the potential to make cryotrons that were more than one thousand times smaller. They could potentially make circuits with connections that were just 0.1 micron in width—that’s 0.001 millimeters. The process was detailed in an undated paper in Buck’s files titled “High Resolution Etching of Evaporated metal Films, K. R. Shoulders, D. A. Buck.”
The manufacturing process was still far from perfect. One of the issues was dealing with the vapors created as a result of the chemical reactions they incited with the electron gun. Some of the chemicals they were using required that a vapor linger for a few seconds to allow the reaction to occur fully. Sometimes these gasses just contaminated the whole experiment.
Buck, Shoulders, and Crawford had understood the basics of making the microchip, however, as the paper explained: “By repeating the process of depositing a metal film, selectively depositing a resist, and then etching, one can form a multilayer structure of narrow conducting paths. Between metallic layers, the same deposition process which is used to selectively form the resist can be used to deposit an insulating layer.”
There was one fundamental difference about the nature of Buck’s work relative to many of the other computing pioneers of the time. Kilby and Noyce were both working on their potential microchips for private enterprise, guarding news of progress in the name of commercial confidentiality.
Buck, the altruistic academic, was sharing details of his work with anyone who would listen. As a famed scientist of the age, he was in constant demand to speak at conferences or to demonstrate the technology to corporations or government departments.
There was still a problem, however, that had already been illustrated by the Arthur D. Little design. No matter how small Buck made his cryotrons, he was reliant on suspending the devices in liquid helium, and traveled around the country with his helium-filled dewar flask—carrying the highly explosive apparatus on commercial flights, trains, and the passenger seat of his car—to demonstrate his latest experiments. Although helium itself is extremely inert, if the connections got clogged with ice, as they were prone to do, the whole thing could blow up.
The Philadelphia computing conference in December 1958 was a milestone moment, however. It was there that Buck announced to the world in granular detail the process that he and Shoulders had devised.
After some deliberation, Shoulders decided against flying from Stanford to attend. Until that point, most of the presentations on manufacturing thin-film cryotrons using an electron beam had been conducted jointly by the two men. Buck was on his own this time.
“The day is rapidly drawing near when digital computers will no longer be made by assembling thousands of individually manufactured parts into plug-in assemblies and then completing their interconnection with back-panel wiring,” he told the conference. “Instead an entire computer or a large part of a computer probably will be made in a single process.”
The paper explained how to deposit chemicals on a surface to form a thin metal film, how to insulate different metals from one another using quartz, and how to repeat the process layer upon layer to make a complete circuit. Buck explained how, in theory, it could be possible to squeeze in fifty million components per square inch, and that it was conceivable to stack ten thousand layers of these components. There was a caveat, however:
This degree of microminiaturization, however, would be difficult to justify in the manufacture of machines as simple as present-day computers. If a computer could be reduced to the size of a cigar box by means of one of the techniques mentioned, there is little point in reducing it further to the size of a postage stamp. The computer would already be dwarfed by its own terminal equipment and, in the case of cryotron circuitry, its Dewar vessel.
Man has ambitious plans for information-handling machines, however, and one can easily envisage a time, a decade from now, when vast numbers of components will be needed in a single machine.
Buck explained how the techniques he was developing could also be adapted to store large documents or to process information. “Such a system might be useful in the telemetering of space-probe photographs back to earth,” he noted. He then talked through a specific experiment that had been conducted by Chuck Crawford in the MIT lab, using films of Parlodion, silicon monoxide, and molybdenum, with masks made from copper mesh. It explained how he let tetraethoxysilane into the vacuum system before firing the electron gun and then added chlorine to make a second chemical reaction. He explained, in other words, how he and his team had already made microchips in their lab experiments. His audience was blown away.
“Zounds!” wrote Ken Shoulders in a letter to Buck a few days after the conference, adding,
From some of the reports I have received on the E.J.C.C. you must have had the after burners on. Some of the fellows came back here days later with their eyes still dilated—to describe your description of the future.
From a purely business point of view, thanks—this may be the kind of nudge that will start our effort here down a fairly nice road. From a personal standpoint, however, it just makes me damn proud to be your associate. Keep your standards high.
Three days after the conference ended, Buck received a letter from Solomon Kullback at the NSA. It was an invitation to join a new panel being assembled by Buck’s friend Louis Ridenour to advise President Eisenhower on electronics and data processing.
As a prelude to taking on this role, Buck was asked to attend a two-day seminar in Washington the following week, related primarily to Project Lightning, the advanced supercomputer program based on his cryotron.
There would be only seven other people on the panel: one from IBM, one from Lockheed, one from General Analysis Corporation, one from Datamatic, another doctor from MIT who worked in the university’s Lincoln Lab, and two from Remington Rand, the company building the Universal Automatic Computer (UNIVAC).
There was little written down on Buck’s letter of invitation to explain what the panel would be about, but the work of its members suggests it clearly tied into the need to build ultrafast computers to support the space race, missile technology, and the continuing work of codebreaking.
After a decade working for the government on secret computer projects as both a full-time agent at Communications Supplemental Activities–Washington and then as a part-time consultant, Buck was being brought deeper into the inner sanctum, with an indirect line to the president. He was, in effect, on the front lines of the Cold War.