The Man Who Invented the Computer

by Jane Smiley

  According to Burton, Atanasoff was floored by this declaration—he had believed Mauchly when he told him at the Naval Ordnance Lab eleven years earlier that the new computer he and Eckert were developing was different from the ABC and “better” than the ABC. But Etienne seemed to know what he was talking about. He said that the particular patent that IBM wanted to challenge was the patent for the memory system—that is, the rotating drum with the rows of capacitors that were regenerated by vacuum tubes. This patent had been finally issued to Remington Rand in the previous year, 1953; possibly IBM knew that Atanasoff had invented this memory system, so this was the patent that they chose to challenge. Etienne asked Atanasoff for all of the relevant paperwork concerning the ABC, but thirteen years, a war, a divorce, and several moves had intervened, and Atanasoff was unable to find what he needed immediately. On June 21, Etienne sent him a copy of Eckert and Mauchly’s patent and Atanasoff wrote back, promising to get him as much of the paperwork as he could. But he never heard from Etienne again, and as far as he knew, the patent challenge was dropped. When Atanasoff read through Eckert and Mauchly’s patent, he saw that it was based on his ideas, but he assumed the case was dropped because the IBM lawyers had decided that breaking the patent was not feasible.

  It was not that IBM had decided that breaking the patent was not feasible; rather, they had decided to make a secret deal with Remington Rand. The deal was to be beneficial for both parties—UNIVAC mostly used an awkward and unfamiliar magnetic tape system for external storage of data; most offices in the market for a computer already had lots of data punched onto IBM cards, and the IBM 650 used these cards. It was not as powerful a computer as UNIVAC, but because of the punch-card storage system, it was a successful entry into the business computer market. But IBM was sued under antitrust laws for using leasing agreements and proprietary punch-card systems to monopolize the office machine market. The solution seemed to be that IBM would sign a consent decree with Remington Rand. The two companies sued each other for patent access. In the meantime, in 1955, Sperry, originally a company specializing in aviation and navigation products (such as gyroscopes, but also the ball turret gun mounted underneath the B-17 during the war), bought Remington Rand.

  About two years after Etienne’s contact with Atanasoff, IBM entered into a private agreement with Sperry Rand, agreeing to pay $10 million over eight years in exchange for access to the ENIAC patents. Once the agreement had been signed, IBM and Sperry aggressively pursued what they considered patent violations by other companies.

  However, IBM and Sperry Rand were busy looking around for other computer ideas. One man whom Konrad Zuse had impressed with the Z4, Helmut Goeze, had married an American woman and moved to the United States. Goeze not only knew that the Z4 could calculate, he knew the amazing tale of its journey from Berlin to Austria at the end of the war. As Zuse writes, “Now in the United States, Goeze wanted to lend his support to this world-important something.” Somehow, Goeze contacted Thomas J. Watson, Sr., who in turn contacted Hollerith Germany, an IBM subsidiary.1 Representatives from Hollerith Germany visited Zuse and the Z4, but they wanted neither the machine nor Zuse’s services—they wanted his intellectual property rights. Over the course of the next year, Zuse negotiated with the company, and as it happened he did make a nice sum of money, in part because the negotiations took so long, and in part because the sum negotiated was in reichsmarks—by the time he cashed the check, reichsmarks had become deutschemarks, which were worth twice as much as reichsmarks. But Hollerith Germany would not hire Zuse for any kind of research—it seems clear in retrospect that throughout the fifties, IBM’s main interest was in cornering the computer market. Zuse did get a research grant from Remington Rand, but it was for a technology that Zuse felt was already superseded—mechanical switching. Zuse thought that he got the grant simply because Remington Rand “still did not completely trust their own [ENIAC-based] electronics, so they wanted to have more than one egg in their basket, just in case.”

  John von Neumann was busy, too, but not on the computer. Once the cold war arms race was well under way, he devoted more and more of his time to advising the United States government, and he gained more and more influence. He may have decided that he had done what he could for computers and that, as Max Newman felt, they were now in the hands of the engineers. And then, in the summer of 1955, he suffered a spontaneous shoulder fracture. That August, he learned that he had a tumor on his left clavicle, probably a metastasis from undiagnosed pancreatic cancer. There was some suspicion that his illness was the result of radiation exposure during his time in A-bomb labs. He was not yet fifty-two. By November 1956 he was in a wheelchair, and by January 1957 he was in and out of the hospital with brain cancer (the Atomic Energy Commission posted a security guard by the door to his hospital room for fear that he would reveal atomic secrets when he was “screaming in horror”). But he continued to advise the government from his deathbed and died on February 8, not quite three years after Turing.

  In early 1959, an IBM official sent an inquiry to Sperry, asking to see the copy of Clifford Berry’s master’s thesis, “Design of Electrical Data Recording and Reading Mechanisms,” which Sperry had obtained in 1953 and in which Berry described the ABC’s regenerative memory. This alerted Sperry, and a Sperry vice president, R. H. Sorensen, began to poke around—he called Iowa State to inquire about the ABC. According to Kirwan Cox, Sperry also hired Howard Aiken to go to Iowa State and look into the matter—he would have, of course, discovered that the ABC had been dismantled. Sorensen took Atanasoff to lunch at the exclusive and elegant Cosmos Club in Washington, D.C. After the lunch, Sorensen sent an in-house memo that conceded that the patents Sperry had inherited from Mauchly and Eckert did overlap with technology already realized in the ABC, but as a result of his meeting with Atanasoff, he doubted that Atanasoff would pursue any legal action—Atanasoff had tried to interest Sorensen in another idea he had for a calculating machine that would have some characteristics of a desktop calculator and some characteristics of a punch-card electronic tabulator. Sorensen politely put him off. Atanasoff was not as gullible as Sorensen thought, however, because after the lunch, he obtained copies of the patents in question, and he saw that they did replicate work that he had done on the ABC. He then went back to his own ordnance business, but not without stowing his new information in a safe place.

  Atanasoff’s Ordnance Engineering Corporation had prospered. In 1956, it was bought for a healthy sum by Aerojet General Corporation, a California company specializing in rocket propulsion technology. Atanasoff took half the proceeds in cash and half in stock—subsequently, the stock split so many times that Atanasoff became a wealthy man. For a few years, Atanasoff worked as vice president and head of the East Coast division, and then, in 1960, he was offered the chance to head the space division, which he turned down. Corporate life did not suit him in several ways—later he said, “I did not want to spend the rest of my life selling and it looked as if the principal effort of the Vice-president of Aerojet was to sell.” Now with plenty of money after a life of frugality, he decided to retire. He was fifty-eight. He immediately embarked upon several projects—he purchased two hundred acres in Maryland and began to design and build an innovative house of a more-than-modern design that incorporated just the sort of unorthodox ideas that a man like Atanasoff would want in his dream house—not only energy-efficient cooling and heating systems and a functional layout, but also tilt-up panel construction and an eight-hundred-pound front door that rotated on brass bearings. He continued to involve himself in the lives of his grandchildren, which could be, according to Burton, less than comfortable for them. She writes, “Retirement mellowed Atanasoff very little, and he remained intense and challenging to others. One reporter described him as ‘creative and cantankerous,’ while his daughter Joanne postulated that ‘conflict was his favorite pastime’ … He enjoyed testing people and was fond of drawing friends and family into intense discussions—or arguments—as a means by wh
ich to grade their mental acuity … he kept tabs on his grandchildren’s schoolwork and carved out time during visits to test us on pertinent material.”

  By 1960, Turing and von Neumann were dead, Arthur Burks was teaching philosophy at the University of Michigan, Max Newman had returned to topology, and Mauchly and Eckert had failed at owning and running their own computer business (though Mauchly had run the UNIVAC division at Sperry until 1959, then started his own consulting firm). Mauchly had received an honorary doctorate from the University of Pennsylvania, the Scott Medal from the Franklin Institute, and other Philadelphia-based awards. Eckert was still with Sperry Rand (he stayed with Sperry, and then Unisys, until 1989). Neither Mauchly nor Eckert had profited directly from the ENIAC patent, but they did get credit (and they did seek that credit) for inventing the computer. Eckert, in particular, was vocal about the inaccuracy of the phrase “von Neumann architecture”—he thought it should be called “Eckert architecture.” But the vagaries of patent law and the delay in awarding the Eckert and Mauchly patents seemed to be working for Sperry. If the patent had been awarded in 1947, it would have run out by 1964, before computers became big business. However, in 1960, the patent was still being challenged. It would not be finally awarded until 1964. At that point, it looked as though it would run into the eighties.

  Zuse finally got to visit the United States and see what computers had been and were being built there, when he and his partner, Harro Stucken, accompanied their mechanical punch calculator test model to Sperry Rand headquarters in Norwalk, Connecticut. Although Zuse understood that the future of computers was electronic, he had contrived a method of doing mathematical operations on punch cards that allowed as many as ten cards to operate simultaneously. It was a mechanical calculator, but it was fast and cleverly conceived, and even though it was never put into mass production, it provided Zuse with funding for his company. Among those they got to visit were General Leslie Groves, who had run the Manhattan Project, and Howard Aiken, who was still advocating using decimal numbers for computers. Zuse writes, “At Harvard they were still completely convinced that the computer was an American invention.” Some years later, Aiken wrote to Zuse, acknowledging the foresight of his earlier ideas. They were also taken to see the Whirlwind at MIT and were most impressed by its size.2 But Zuse’s business connections were Swiss more than American, and eventually the Z4, after years in a barn in the Austrian Alps, and thanks to the man in the elegant automobile, it was sent to Zurich, “the sixth transport we put it through.” When the day came to demonstrate it, the Z4 started sparking and then went dead during an afternoon test run. Zuse and his partners did not panic, though—they discovered that the problem had to do with a newly installed transformer and fixed it: “We had exactly a half an hour to correct the error and replace the burned out lines. We did it, aired out the faint burning smell, and at four o’clock our illustrious guests witnessed a perfect demonstration.” Eventually, Zuse came to have his “fondest memories” of his years in Zurich. He admired his colleagues, and his computer continued to operate so reliably that it could be left on, unattended, overnight. He writes, “Many a night, I walked through the lonely streets of Zurich, on my way to check on the Z4. It was a strange feeling, entering the deserted ETH3 and hearing, already by the time I reached the first floor, that, on the top floor, the Z4 was still running perfectly. In those days you could tell from the rhythm of the punched tape reader.”

  In 1962, Richard Kohler Richards, who had a doctorate in electrical engineering, had worked at IBM, and had written several books on computers including Arithmetic Operations in Digital Computers and Digital Computer Components and Circuits, decided to return to Ames, where he had been an undergraduate at Iowa State, and write a book about the history of the computer. His neighbor turned out to be a man named Harry Burrell, who remembered writing a press release about the Atanasoff-Berry Computer around the time that the Des Moines Tribune ran a brief article, with a picture, about the machine (January 15, 1941). The article stated, “An electrical computing machine said here to operate more like the human brain than any other such machine known to exist is being built by Dr. John V. Atanasoff, Iowa State College Physics Professor. The machine contains more than 300 vacuum tubes and will be used to compute complicated algebraic equations. Dr. Atanasoff said it will occupy about as much space as a large office desk. The instrument will be entirely electrical and will be used in research experiments.” But there was no record of or paperwork concerning the machine in either the library or the engineering publications office. It was then that Richards visited Sam Legvold, who had returned to the physics department at Iowa State after the war and had worked with Atanasoff on his defense department project in the basement of the physics building, right next to the ABC, and later with him at the NOL.

  Legvold remembered the ABC quite well, and not only that, he had a drum from the computer that he had salvaged from the 1948 wreckage. He also remembered talking with Berry about the computer, though not with Atanasoff—with Atanasoff, he had only discussed the defense project they were working on. Legvold was not the only physics professor who remembered the ABC, but no one remembered how it worked (if they had ever known) or the principles behind it. In February 1963, Richards wrote to Atanasoff to inquire about the machine, but Atanasoff was too busy with his retirement projects to give him much help. Once again he was moving house—this time building the house—and once again, perhaps, the paperwork didn’t seem worth finding. Atanasoff always invested himself fully in his project of the minute, and in addition, none of his contacts with IBM or Sperry about the ABC had ever come to anything. He suggested that Richards contact Clifford Berry, who was younger and might remember the ABC in more detail.

  In March, Richards wrote to Berry. Berry was now in his early forties, still married to Atanasoff’s former secretary, and gainfully employed in the research and development department at Consolidated Engineering Corporation (later to become a part of Bell and Howell and then DuPont). Consolidated Engineering specialized in developing mass spectrometers. In 1945, Berry had invented his own small computer for the purpose of sorting through the large amount of data produced by the mass spectrometer. Berry had invented many other things—eventually, he owned almost thirty patents in addition to the patent for his small computer. Richards also wrote to the UNIVAC division at Sperry, looking for John Mauchly’s address.

  Berry replied ten days later. He remembered the ABC perfectly well. He directed Richards to his master’s thesis in the Iowa State library and also told him about the report for the Iowa State College Research Corporation and the patent applications that had been written but never filed. He added, “An interesting sidelight is that in 1940 or 1941 we had a visit from Dr. John Mauchly who spent a week learning all of the details of our computer and the philosophy of its design. He was the only person outside of the Research Corporation and the patent counsel who was given this opportunity, and he may still have notes of what he learned from us.” Berry then went on to give a concise description of the ABC. He wrote:

  I am not sure what Dr. Atanasoff told you about the machine so I will describe it briefly. The machine was designed specifically to solve sets of linear simultaneous algebraic equations up to 30 × 30. All internal operations were carried on in binary arithmetic; the size of the numbers handled was up to 50 binary places (about 15 decimal places). Initial input of data was by means of standard IBM cards, with five 15-place numbers per card; the machine translated the numbers to binary numbers. The machine’s “memory” consisted of two rotating drums filled with small capacitors. The polarity of the charge on a given capacitor represented the binary digit standing in that position. A “clock” frequency of 60 cycles per second was used, the mechanical parts of the machine being driven with a synchronous motor. Storage of intermediate results was by means of a special binary card punch, with which 30 binary numbers, each 50 digits long, could be punched on one card. The mathematical method employed to solve sets of equat
ions was that of systematic elimination of coefficients through linear combinations of pairs of equations.

  He included six pictures as well as copies of the news stories about the ABC. For the next few months, Richards and Berry conducted a detailed correspondence about the ABC. Berry, still in the computer business, was amazed to discover that the record of the ABC at Iowa State was so thin, and also that Atanasoff himself had not kept up with what was going on in computers sufficiently to maintain the record of his own contributions. The correspondence supplied Richards with enough detailed information to establish apparent links between the ABC and ENIAC.

  Mauchly did not respond to Richards’s first letter and then did not return his calls. But Richards was persistent. When he finally reached Mauchly in the late summer, Mauchly was not happy to hear from him. He derided the ABC, but he did admit to staying in Ames for several days, looking at the computer, and discussing it with Atanasoff. Richards later wrote in his book Electronic Digital Systems, “The Atanasoff Berry computer … does … appear to predate every other electronic digital system by a matter of years.”