The Man Who Invented the Computer

by Jane Smiley

  In the meantime, Berry decided to take another job, this time in Huntington, New York, on the North Shore of Long Island, at the Vacuum-Electronics Corporation. He left Pasadena in early October and went to Long Island, stopping at a conference for a week on the way. He rented a room, intending to look for a house (he found two), buy a new car, and prepare the way for his family to move east. The company agreed to his request to bring Jean Berry east to New York so that she could choose between the two houses. Berry called Jean every night, and he seemed to her to be excited about both his new job and their new life. But on October 30, before she was due to leave for New York (on November 6), she received a phone call from the Huntington police—Clifford Berry had been found in his rented room, dead, with a plastic bag over his head. The cause of death was listed as “probable suicide.”

  Jean Berry discovered when she went east that the police were not sure of what had happened—they maintained a sealed-off crime scene in the room where the death occurred for three weeks. Atanasoff was suspicious enough to drive from Maryland to Long Island and talk to the landlord, who declared that he himself had easily removed the plastic bag from Berry’s face. Jean Berry and Atanasoff eventually became convinced that Berry had not committed suicide, though Scott McCartney raises doubts about Berry’s mental condition in his defense of Eckert and Mauchly, by stating that he had been in two car accidents, which “left him in substantial pain,” and that “he was intoxicated at the time of his death.” In fact, when Jean Berry first told Atanasoff about Clifford Berry’s death, she did not mention suicide at all, but said she thought that it might be related to head injuries suffered in a car accident in 1956 that had resulted in occasional seizures. But Jean Berry later wrote, “When I told a physician what I knew, he said that Cliff could not have possibly killed himself—he was murdered: ‘It’s like trying to hold your breath; you can’t.’ ” She believed to the end of her life that he had indeed been murdered (though there is no public record of who she thought was responsible). Others shared her belief. Kirwan Cox, the Canadian filmmaker who has researched Atanasoff and Berry and done numerous interviews, maintains that whether Berry was or wasn’t murdered, the unarguable result of Berry’s death was its “huge impact on Atanasoff. Prior to Berry’s death, Atanasoff had not wanted to discuss the ABC, because he was too upset about the destruction of the ABC. But Atanasoff believed Berry was murdered, and that he would not have died if Atanasoff had not hired him to work on the machine. [The] death of Berry changed his attitude to the patent lawsuits, and he became quite energetic in pursuing the patent conflict.”

  When R. K. Richards’s book Electronic Digital Systems was published three years later, in 1966, some of the first people to read it were patent lawyers at several computer companies. One of these was a man named Allen Kirkpatrick, who had been hired by Control Data Corporation (the home of Seymour Cray, who was later to found Cray Research) to defend CDC against a case of patent infringement brought by Sperry Rand. CDC was being sued along with Honeywell, and they had decided to collaborate on their defense. Richards’s book was sizable and respectable, given his earlier work. It was Richards who coined the term “Atanasoff-Berry Computer,” and in the book’s preface, he stated point-blank, “The ancestry of all electronic digital systems appears to be traceable to a computer which will be called the Atanasoff-Berry Computer.” Since in spite of the lingering patent controversies ENIAC was famous for being the world’s first computer, it proved something of a shock to the computer world when Richards stated, “There was, however, one interesting link between the machine and later work. One of the few people to study the [ABC] in detail was Dr. John Mauchly … According to oral reports from Dr. Atanasoff and Dr. Mauchly, the two met at an American Association for the Advancement of Science Meeting. As a result of conversations at this meeting, Dr. Mauchly made a visit to ISU in 1941 for the specific purpose of studying the computer. As mentioned later, Dr. Mauchly is given credit for subsequently initiating the ENIAC project.”

  Control Data Corporation may have had a special desire to break the Eckert-Mauchly patents because they had firsthand knowledge of them. CDC had originally been a company called Engineering Research Associates and had grown out of a World War II U.S. Navy code-breaking operation. Just after the war, ERA continued to build code-breaking machines designed around rotating drums and paper tape readers (there is no evidence that the ERA inventors knew about Colossus), and they did successfully break several Soviet codes, but when in 1949 the Soviets changed the code that had been broken the previous year (shades of what had happened with Enigma in 1942), the machine they had devised stopped being useful. ERA at first decided to go into scientific computers, but then there was a conflict-of-interest scandal on the military procurement side, and the company went broke and was sold to Remington Rand. In the mid-fifties, when Remington Rand was bought by Sperry, the ERA computer group was consolidated into the UNIVAC division. The original ERA group grew restive at UNIVAC and left to form Control Data. They did well—by 1964, the Control Data CDC 6600 supercomputer had successfully challenged the comparable IBM computer, especially in terms of processing speed (three times faster than the IBM). CDC was hard at work on the next version; the stakes were high, and Sperry Rand, IBM, and Control Data knew it.

  Honeywell was a much older company, owing its existence to the 1885 invention of a thermostat for coal furnaces. By the 1960s, Honeywell’s technological products had ranged away from heating and plumbing inventions into all sorts of other fields, including the autopilot mechanism for aircraft, which was invented during the war, the ubiquitous round wall-mounted thermostat, and many sorts of gyroscopes. Honeywell got into the computer business by joining with Raytheon to form Datamatic and then buying out Raytheon (Raytheon was the company founded in 1922 by Vannevar Bush, inventor of the Bush Analyzer).

  CDC and Honeywell were beginning the suit at a disadvantage—Bell Labs, where George Stibitz had invented his K-for-Kitchen calculator in 1937, had already tried suing Sperry and lost. The judge said that Bell had not produced evidence of “prior public use” of the ideas incorporated into ENIAC. The lawyer assigned to the Honeywell/CDC case knew this because he had worked at the law firm that pursued the Bell Labs case. Before Richards’s revelations, the CDC/Honeywell defense focused on the competing claims of engineers and scientists who had worked with Mauchly and Eckert on ENIAC—plenty of them felt that in their broad patent, Mauchly and Eckert claimed ideas that other people had come up with. Because of this, Honeywell and CDC hoped that Sperry might be willing to negotiate, but they couldn’t count on such an eventuality, and even while proposing a settlement, they began working on a different approach.

  Honeywell and CDC had several pieces of luck—one of these was that the general counsel of the patent division at Honeywell was an Iowa State College graduate in electrical engineering, and a classmate of R. K. Richards. Allen Kirkpatrick and his assistant, Kevin Joyce, were also electrical engineering graduates who had gone on to law school. When they read what Richards wrote about the ABC, they understood it, and when they visited Atanasoff, always prickly and impatient with the ignorant, they could talk to him and convince him that they understood what he was saying. Perhaps their greatest piece of luck, though, was that when Atanasoff at last took the time to rummage through all of the old boxes he had been moving over the twenty years since leaving the house on Woodland Street in Ames, he found everything he had kept, and everything they needed.

  The most important member of the legal team was a young lawyer named Charles Call, who in 1966 was twenty-eight years old and had already worked on six successful patents for Bell Labs. Call was familiar with vacuum tubes and ham radios. He understood the Richards book, and he was able to understand two other documents that he obtained from Iowa State—Clifford Berry’s master’s thesis and Atanasoff’s thirty-five-page description of the ABC, written in August 1940. By the time he read the Atanasoff documents, he had done considerable work on the case already,
and, as Clark Mollenhoff points out, “Studying Atanasoff’s memorandum against the background of his months of study of the ENIAC, EDVAC, and UNIVAC patents, Charles Call became convinced that Atanasoff’s concepts at the time of Mauchly’s visit were far ahead of his time. Also, they went beyond ENIAC and included many of the most important concepts of such second-generation … computers as EDVAC.” As Kirwan Cox notes, in contradiction to Mauchly’s remarks to Atanasoff during his visits to the Naval Ordnance Lab, progress toward the modern computer involved adhering more closely to the ABC model, not moving away from it.

  But Call knew that Honeywell and CDC were still at a disadvantage—as good as the documentation looked, it would be difficult to establish to the satisfaction of a judge that Atanasoff’s claim to prior art (something publically known or published about an invention that challenges that invention’s claim to novelty or “nonobviousness”) was more important than leaving things as they were. And Atanasoff himself was now in his mid-sixties—though he looked healthy, with Berry dead, he was the only source for detailed technical information. To safeguard this aspect, Call videotaped Atanasoff’s depositions (a first, according to Tammara Burton) and photographed every page of his documentation. It was only after reading through these copies that Call began to feel confidence in the Honeywell/CDC case.

  There were in fact two cases—the Honeywell case concerned the ENIAC patents, which covered more than a hundred ideas (after Mauchly and Eckert lost possession of the EDVAC ideas in 1947, they had decided to claim as much ground as possible). The CDC case covered only one patent, patent 827, concerning what Atanasoff had called “regenerative memory”—this was the same patent IBM had proposed challenging in 1954. The Honeywell case had a rather dramatic beginning—on May 26, 1967, as soon as the Sperry lawyers signaled the company’s unwillingness to settle, a runner from the Minneapolis law firm hurried to the courthouse to file the case. He arrived there fifteen minutes before his counterpart in Washington, D.C., arrived at his local courthouse. This, plus a subsequent finding by Judge John Sirica, in Washington, that the case would take up too much time in the crowded District of Columbia schedule, meant that the case was to be tried in Minneapolis. The less dramatic CDC case was to be tried in Baltimore.

  Sperry Rand, in the meantime, wasn’t focusing only on CDC and Honeywell—the company was also suing General Electric for patent infringement. One of the attorneys for General Electric turned out to be George Eltgroth, who had helped Eckert and Mauchly file the original patents and connect with American Totalizer (another of the GE attorneys was an electrical engineer who had helped Berry build the ABC while at Iowa State). Eltgroth had never heard of Atanasoff, Berry, or Mauchly’s trip to Ames in 1941. This meant that if a connection could be proven, and Mauchly had knowingly withheld that information, he would have failed to comply with full-disclosure rules for patents. In this context, Eckert’s October 1953 remarks about the ABC in the Journal of the Institute of Radio Engineers that had originally alerted IBM to a potential patent problem were also significant. Eltgroth heard about Atanasoff in a meeting devoted to GE defense. He exclaimed, “If I had known, I could have protected them!” Why Mauchly had acted as he did, and indeed how his mind worked, subsequently became a matter of considerable interest and deepening mystery.

  Another piece of luck for Honeywell and Control Data was that on the very day when Call finished reading over his copies of Atanasoff’s documents, including Mauchly’s enthusiastic letters to Atanasoff after seeing the ABC, he happened to go to a panel on computer science chaired by a man named Isaac Auerbach, who had, in 1960, established the International Federation for Information Processing Societies. Mauchly was to be on the panel and was listed in the program as the “inventor of the first automatic electronic digital computer.” Auerbach, who had worked on ENIAC and had been employed at Sperry UNIVAC, had also read Richards’s book. He asked Mauchly to comment on Richards’s assertions about the Atanasoff-Berry Computer. Mauchly admitted that he had gone to Ames to see the computer and that he had talked to Atanasoff about it. Then he gave Call a foretaste of his future testimony—the computer hadn’t worked, he hadn’t learned anything from Atanasoff, he hadn’t spent much time with the machine. Since Call had Atanasoff’s letters and documents, all of which corroborated an entirely different story, Call knew that such a defense would hurt Sperry’s case, whatever Mauchly’s motives. Whether he remembered what had really happened and was banking on Atanasoff not retaining the documents, or whether he actually had no memory of his response to the ABC, he would be seriously compromised either way.

  1. For more information about this connection, see Edwin Black’s IBM and the Holocaust: The Strategic Alliance Between Nazi Germany and America’s Most Powerful Corporation (New York: Crown, 2001).

  2. The Whirlwind was another offspring of ENIAC. A man named Perry Crawford was working at MIT, trying to create computerized flight simulators for the navy. They were using analog ideas before Crawford saw ENIAC in 1945. Subsequently, the U.S. Air Force based the SAGE early warning system on the Whirlwind.

  3. Eidgenössische Technische Hochschule Zürich.

  Chapter Eleven

  Before the death of Clifford Berry, Atanasoff had been reluctant to involve himself in the patent dispute between Sperry Rand on one side and Honeywell and Control Data on the other because the only thing in it for him was recognition, and that was uncertain because as far as the courts were concerned, it meant returning to a question that had already been decided in favor of Sperry Rand and it meant abrogating patents that had long been in dispute and then issued by the U.S. Patent Office. The case therefore involved at least an implicit challenge to the patenting process itself. For that reason, Honeywell and CDC had a small advantage in the fact that the case would be tried in Minneapolis rather than in Washington, D.C. But all the lawyers for Honeywell and CDC knew it was an uphill fight, and Atanasoff did too, because he had studied patent law with his usual energy in the course of his business ventures.

  Once he was engaged, Atanasoff took his customary pedagogical position and tested the lawyers to see if he could count on them to understand the ideas behind the ABC. But there was more to it—it was as if he could only participate wholeheartedly if he could thoroughly understand a process and a system. He had to learn everything he could about it in order to get it into his mind and go forward—just as he had taught himself to drive the family Ford at age twelve by learning everything he could about how the automobile worked and how to fix it if something went wrong. And so he used the lawyers to learn what he needed to know. Probably they privately considered the old man a pest.

  Atanasoff made a list of witnesses to be interviewed and deposed. On the list were those whom he remembered as having been around during the construction of the machine and during Mauchly’s visit to Ames—notably absent, and profoundly missed, of course, was Clifford Berry. Prominent on the list was Robert Mather, a professor at Berkeley who had worked on the computer with Berry as an undergraduate at Iowa State. But Mather, who had not been around for Mauchly’s visit, told Call and Kirkpatrick, “I just wasn’t sophisticated enough to particularly notice who [people who visited the machine] were.” He also said, “You see, Cliff did most of … the more complicated things, the more routine things were turned over to me. I was soldering the wires to brushes and the terminal board to the binary-to-decimal converter.” An interesting addendum to this quotation from Mollenhoff’s book is the observation by John Gustafson that “I’m not sure we could have reconstructed the ABC without Mather’s input. He was proud of the fine job he had done of wiring the machine neatly, and it was one reason he took those sharp black-and-white photos of the computer. I’m not sure we have a single photo of the original ABC that was not taken by Mather.”

  A more productive interview was conducted with Sam Legvold. Though he had never talked to Atanasoff about the computer, he had been very good friends with Clifford Berry. He also had both a strong interest in the ABC and
an excellent memory. A new graduate student in the fall of 1939, Legvold remembered the pre-ABC prototype in considerable detail—he had seen it operate, and he also remembered many things Berry had told him about how it worked and how it was constructed. He remembered, too, visits by experts—a representative from the Rockefeller Foundation and a man from MIT and the NDRC. He also remembered Mauchly, and in some detail.

  Legvold, about twenty-two at the time of Mauchly’s visit, remembered going to lunch with the thirty-three-year-old Ursinus professor and finding him “a rather delightful fellow, pretty bright and stimulating.” He remembered him “being in there with his shirtsleeves rolled up, pitching in to help do some things on the computer as we sat and talked about it.” In the course of his own work, Legvold passed through the computer room quite frequently. He remembered that Mauchly had been around for three days—“more than just a drop-in-for-an-afternoon-kind of thing.” He also remembered that Mauchly had taken “a sharp interest” in the ABC; the discussions among Berry, Atanasoff, and Mauchly had been “free and open,” and he had seemed to understand the principles behind the machine.

  Once Charles Call (who was representing Honeywell) had finished with Legvold, it was time for Allen Kirkpatrick, who was representing Control Data, to interview Mauchly. Kirkpatrick was canny and unrevealing in his questioning, and Mauchly, though under oath, seemed naive, or guileless, in his answers. (Scott McCartney says his memory was bad because he was in poor health.) He acknowledged that he had written to Atanasoff in the six months after his visit to Ames, and that his letters had been enthusiastic. He acknowledged that he had asked Atanasoff about building “an Atanasoff calculator” at the Moore School, but he became confused (or “flustered” as Mollenhoff says) when asked to explain the questions in his letter a little more clearly (neither his memory nor his records were as good as those of Atanasoff and his associates). When shown the October letter, he said, “The center portion of this letter indicates that I was probing whether there would be any objection to using some of his [Atanasoff’s] ideas. This is not quite as strong as saying that I had a strong desire to, but at that point, on September 30, 1941, I think the letter makes it clear that I was still seeking a good way of implementing an electronic calculator, and this is the same interest which I displayed with respect to many other ideas with respect to computation, such as those which I saw at the World’s Fair in 1939.”