by Kurt W Beyer
Even Aiken’s uncompromising insistence on reliability and punctuality interfered at times with the design process. “Howard got to the point where he says, ‘You know, on April 1st we’re going to stop designing and build it anything from that time on will go into the book for our next machine,’ ” Miller described. “And sure enough, if somebody tried to add a feature he would stop them. I recall somebody trying to put something in on the next day and he said, ‘No. That’s the date. We’re going to build it.’ ”47 The same virtue of discipline that permitted Aiken to succeed where less focused inventors like Charles Babbage failed, had become a vice that interfered with the ability of the Computation Laboratory to explore more exotic technologies.
The Harvard Computation Laboratory at night. Courtesy of Harvard University Archives.
THE BEGINNING OF THE END: THE MIGRATION OF TALENT
In the days that followed the first Symposium on Large-Scale Digital Calculating Machinery, the Christian Science Monitor, the New York Times, and other newspapers praised the accomplishments of Howard Hathaway Aiken and his staff. Aiken once again took full advantage of the wide reach of the Harvard News Bureau to inform the world of their accomplishments. And once again, Grace Hopper found herself adorning the cover of a major newspaper. This time a well-dressed Hopper is pictured hobnobbing with Navy brass, while another photograph had her reviewing printouts from the Harvard Mark I.48 Aiken also took full advantage of Richard Babbage’s visit, and the two appeared in a photograph together inspecting the gears of Babbage’s Difference Engine.
In the eyes of the educated public, Aiken, Hopper, and the rest of the Harvard Laboratory crew were situated at the forefront of a technical revolution the likes of which were only beginning to be understood. They had designed, constructed, and operated two machines before other groups had assembled one. It was only prudent that others should travel to Harvard and learn from the intellectual heir of Charles Babbage and his disciples.
For the members of the Harvard crew, the Symposium exposed them to a new world of ideas, many of which contradicted much that they had learned while working with Aiken. The Commander had created a closed system cut off from the ideas of others. “The Harvard project was sort of self-contained,” Campbell recalled. “It didn’t—at least through 1950—it pretty much went its own way and there was not an awful lot of inter-change.”49 Aiken justified the self-containment on the grounds that Harvard had a monopoly on good ideas. “You worked with Howard,” Miller recalled, “you knew everything, you had the world in your fist, and everybody else must be way out of line and out of base. In fact, I have a funny feeling that if I’d asked to spend some time at MIT to see what they were doing, he’d have seen to it that I got fired some way.”50
But the Symposium had opened up the floodgates, and for some of the crew, there was no turning back. Weeks after the Symposium, Robert Campbell left Harvard to join Raytheon Manufacturing Company. The company had decided to set up a computer group, headed by Andy Adelson, formerly of the MIT Radiation Lab. Besides the financial incentive, Campbell would be in charge of systems design with wide-ranging authority.51 What shocked Campbell most upon arriving at Raytheon was the collegial environment among his co-workers. “At Harvard,” he recalled, “the whole situation was dominated by Aiken. At Raytheon it was more the group of us working together.”52
After completing a study for the Bureau of Standards, Raytheon won a contract to build a computer for the Naval Missile Test Center at Point Mugu, California. The computer, designated RAYDAC, would be a binary, electronic computer with a stored program, magnetic tape input and output, and mercury delay line memory. Before the ink on the RAYDAC contract was dry, Campbell convinced Richard Bloch to join him at Raytheon. “Aiken felt that I really should get into the professional ranks there [at Harvard] and remain in the academic environment,” Bloch recalled. “I tended to feel that I was destined to move into industry and that there may be greater challenges there.”53
Bloch left the Harvard Computation Laboratory in March 1947 to head up the software side of the RAYDAC project. Bloch basked in his newfound intellectual freedom, and soon made significant contributions to the Raytheon project. “The thing I was proudest of was the fact that I came up with the automatic error detection system,” he recalled.54 Bloch’s error detection system was the first instance of an extended parity check. When Robert Campbell took a more senior position at Burroughs in 1949, Bloch was placed in charge of both hardware and software design.
THE FORMATION OF THE ASSOCIATION FOR COMPUTING MACHINERY
According to Hopper, the cross-fertilization and heightened communication generated at the Harvard Symposium made many attendees aware that the emerging computing community lacked the necessary institutions to permit a broad dialogue to continue beyond Harvard.55 In particular, Edmund Berkeley, a former Computation Laboratory crew member, believed that a forum for sharing knowledge was needed for all those interested in computers, not just a select few. “We felt that we were not so worthwhile in the eyes of people who were making arbitrary decisions about what’s worthwhile,” recalled Harry Goheen, another former Harvard crew member and early supporter of Berkeley’s vision.56 Edmund Berkeley was certain that he was worthwhile and was willing to do something about it.
Berkeley received his bachelor’s degree in mathematics from Harvard University in 1930. Upon graduation he worked for New York Mutual Life Insurance and then moved to Prudential Insurance. After the attack on Pearl Harbor, Berkeley joined the Navy and was assigned to the Dahlgren Ordnance Laboratory as a mathematician and a human computer. When the Bureau of Ordnance contracted with Howard Aiken to build Mark II, Lieutenant Commander Berkeley was sent to the Harvard Computation Laboratory to monitor the development of the project.
From the start, Berkeley had a difficult time adjusting to the Computation Laboratory’s environment. Despite his rank, Hopper and the crew saw him as a civilian computer. “He didn’t know much about the Navy,” Hopper recalled. “He got scared to death of Aiken so he started sending me little notes.”57 When Hopper got tired of serving as an intermediary and Berkeley finally gained the courage to face Aiken, he exhibited a nervous habit of jotting down everything the commander was saying instead of looking at him. “This drove Aiken crazy,” Hopper recalled, “so Aiken wouldn’t let him take notes, and so he would go out to his desk and make notes as soon as he finished talking to the commander.”58
Berkeley’s shy disposition and penchant for order did not go unnoticed by the rest of the crew. “We used to plague the hell out of Berkeley,” Hopper recalled. The crew’s favorite activity was stealing Berkeley’s trusty date stamp. Berkeley date-stamped every document that came across his desk, much to the amusement of some of the crew. “One day they very carefully stole his date stamp, and went up and date-stamped the whole roll of toilet paper in the men’s room,” Hopper recalled. “Berkeley threw a fit when he found his precious date stamp.”59
Practical jokes aside, both Berkeley and his co-worker Harry Goheen eventually became fed up with the Computation Laboratory’s “repressive” atmosphere and went their separate ways.60 Berkeley was discharged from the Navy and returned to Prudential Insurance as the chief research consultant. During the Harvard Symposium in 1947 Berkeley and Goheen were reunited with Robert Campbell and the three began making plans for a forum for interchange between individuals and organizations at work in the computing field. “He [Berkeley] and several of us had felt that since Aiken handled most of the external contacts himself , other people in the Computation Laboratory were isolated from the rest of the world,” Campbell recalled. “That was one of the motivating influences in the founding of ACM. Ed Berkeley, Harry Goheen and I from Harvard with several other people helped get the thing organized.” According to Campbell, Berkeley served as the “principal instigator.”61
It took Berkeley about 4 months after the Harvard Symposium to fashion a proposal for the new organization, at which time he attempted to enlist the
support of the established members of the new field. “One day in April, 1947,” Berkeley recalled, “I went to see Archibald and Lehmer on the evening before a meeting of the National Research Council Committee on Large Scale Calculation. I persuaded Lehmer and Archibald to take the matter up at the meeting the next day, in which Stibitz and von Neumann and Aiken were . . . to meet with them.”62 Berkeley’s proposal was brought up, met significant resistance, and voted down.
Howard Aiken remembered the Berkeley proposal and his strong feelings against it: “I said, ‘No, we shouldn’t do that because computation was a universal thing.’ That our best interest and the best interests of the scientific community as a whole would be better served to assist everybody to use machinery and to publish their work, . . . von Neumann agreed with that completely, and so this proposal of Curtiss was voted down.”63 Aiken mentioned John Curtiss because he was the most ardent supporter of the Berkeley proposal on the committee. Curtiss, who at the time was the division chief of the National Bureau of Standards’ Applied Mathematics Division, broke ranks with Aiken and von Neumann and joined Berkeley’s quest for a computing association. That summer, Berkeley and Curtiss organized a temporary planning committee, which included Robert Campbell, Dick Bloch, and Harry Goheen. The group approached John Mauchly who, along with Presper Eckert, had recently formed the first private computer company, Electronic Control Company. Initially, Mauchly did not see the need for the organization. “John Mauchly said that there was complete freedom of interchange of information,” remembered Goheen. “Well hell, he was a director of the project that built the ENIAC. Of course, for him there was complete freedom, and for people like von Neumann and Aiken—I’m sure there was never any difficulty finding out what was going on.”64 Eventually the group convinced Mauchly of the organization’s potential utility, and the date was set for the initial gathering.
The first meeting of the Eastern Association for Computing Machinery took place on 15 September 1947 in a small physics laboratory at Columbia University. The majority of the meeting was spent on administrative issues, although T. Kite Sharpless reported on the design of the University of Pennsylvania’s EDVAC computer. Foremost on the agenda was the selection of the Executive Council, which proved more difficult than anticipated. “There were,” Goheen recalled, “people from IBM that felt that this was a ploy . . . of some of Aiken’s friends to form another self-adulation society for the benefit of non-IBM people, and they were opposed to that. I mean, they didn’t realize that we were a bunch of outs, I’m sure. But their attitude was: these are some of Aiken’s boys.” 65 According to Goheen, it took considerable reassurances on the part of Berkeley to placate the IBM crowd, and eventually all sides called for a vote. The 78 people in attendance elected Edmund Berkeley as (Prudential Insurance) acting secretary and Robert Campbell (Raytheon) as treasurer. The two most prominent figures at the meeting, John Mauchly (Electronic Control Company) and John Curtiss (Bureau of Standards), were elected vice president and president respectively. Other appointments to the executive council by the elected officials included Mina Rees (Office of Naval Research), C. B. Tompkins (Engineering Research Associates), Richard Taylor (MIT), T. Kite Sharpless (Technitrol, Inc.), E. G. Andrews (Bell Laboratories), and Franz Alt (Ballistic Research Laboratory).66
The assembled group agreed that the purpose of the organization was to “advance the science, development, construction, and application of the new machinery for computing, reasoning, and other handling of information.” Initially, the organization would represent four geographic areas—Boston, New York, Philadelphia, and Washington—with each elected official representing one of the four chief geographical sections. The founding members agreed that the “Eastern” would be removed if and when the association had ample representation from the Mid-West and West Coast.67
National aspirations aside, the fact remained that other than John Mauchly and John Curtiss, the majority of the computing elite were conspicuously absent from the meeting, above all Howard Hathaway Aiken and John von Neumann. “It would have been so much better,” Berkeley recalled, “if we could have gotten Howard Aiken, and John von Neumann, and Herman Goldstine, and a few other people to work with us in the society in the early days. It wasn’t through a lack of trying.”68 Berkeley, in fact, sent von Neumann a letter one month before the meeting encouraging the Princeton professor to participate. Von Neumann responded in the negative, reminding Berkeley that although he felt that in theory a computing association was desirable, he believed that the overall situation had not matured sufficiently to support such an endeavor. Despite his personal views, von Neumann did end the letter on a positive note, indicating that “this does not mean, however, that I will not be very glad if you succeed in furnishing the proof of the opposite, and I want to use this occasion to wish you the best of luck in your efforts.”69
Berkeley, through his persistence, was able to convince von Neumann to present a paper at the Association’s first general meeting held 11 and 12 December 1947, at the Ballistics Research Laboratories, Aberdeen Proving Ground, Aberdeen, Maryland. Von Neumann accepted the invitation, but Goheen noted, “he was willing to appear, but as von Neumann, not as a lecturer for the ACM.”70 The famed mathematician’s paper, titled “General Principles of Coding, with Application to the ENIAC,” was the first example of a conference paper dedicated to programming rather than machine design.71
Von Neumann’s presence at the Aberdeen meeting helped to legitimate the nascent organization, and by January 1948 membership swelled to 350. On 16 January 1948 the Executive Committee dropped the “Eastern” and changed the name to Association for Computing Machinery. Moreover, Curtiss arranged for the National Research Council journal, Mathematical Tables and Other Aids to Computation, to reserve a section for articles and papers generated by ACM members. But despite the success of the former members of the Harvard Computation Laboratory just one year after the Harvard Symposium, nothing could change the mind of their former boss. “Aiken,” Campbell recalled, “was always his own man. He went his own way and he had his own idea of how to do things.”72 Eventually Campbell had to trick his former boss into joining the organization, “I asked him when he was going to pay me back the five dollars he owed me. He didn’t remember, but he paid me back the five dollars. I used the five dollars to enroll him in the ACM. I don’t think he ever paid another dues.”73
HOPPER’S PLACE IN THE COMPUTING COMMUNITY
While former Harvard crew members Edward Berkeley, Robert Campbell, Harry Goheen, and Richard Bloch were off creating the Association for Computing Machinery, Grace Hopper was back at Harvard managing the operation of Mark I, training personnel from Dahlgren Ordnance Laboratory, and entertaining and educating visitors to the Computation Laboratory. Once again, Aiken took advantage of Hopper’s writing ability and had his diligent assistant write the manual for the Harvard Mark II; it was published by the Harvard University Press in 1949, again with “Staff of the Harvard Computation Laboratory” rather than “Grace Hopper” as the author.
Despite the apparent injustice, Hopper was quite content, even thrilled, with her place in life. She was skilled at programming and had full access to not one but two of the only functioning large-scale calculating machines in existence. She had a passion for teaching anyone who cared to learn about her machines, and the Computation Laboratory provided an endless stream of visitors and students. Finally, Hopper, more than most, had a special bond with Aiken, especially after the departure of Campbell and Bloch. Hopper enjoyed Aiken as a person, and respected him as a boss. In turn, Aiken bestowed on Hopper unparalleled liberties in decision making, and a healthy dose of responsibility for the day-to-day operation of the Computation Laboratory. Within Aiken’s microcosm, Hopper was a somebody.
Hopper’s hard work and dedication had convinced a once-skeptical Aiken that a woman could do what he considered to be a man’s job, but eventually Hopper could not overcome the immovable rules of Harvard University. With the expiration of
the Navy contract in 1949, Aiken continued on as a tenure track professor at Harvard. He remained director of the Computation Laboratory until his retirement in 1961.74 Hopper, on the other hand, was a faculty research fellow with a three-year contract. “They didn’t promote women at Harvard at that point, so at the end of three years my time was up,” she recalled.75
Fortunately for Hopper, by the start of 1949 the Association for Computing Machinery had taken root as the primary organization for the emerging computing community. Hopper’s former co-workers Robert Campbell and Edmund Berkeley were reelected in January to serve as treasurer and secretary respectively, and the next significant meeting was planned for 18–20 April at the Oak Ridge National Laboratory in Tennessee. Hopper traveled to Oak Ridge to reunite with old friends, but also to test the job market. The talented programmer spoke with the heads of a variety of computing projects and computer-related organizations and was overwhelmed by the number of job offers she received.76
On the back of her conference program, Hopper jotted down her career options. She received offers from the Ballistics Research Laboratories, Aberdeen Proving Ground, Engineering Research Associates, the Eckert-Mauchly Computer Corporation, and two separate offers from the Office of Naval Research. Both Engineering Research Associates and the Eckert and Mauchly offered the most money, a substantial salary of $7,200 per year. Next to one of the ONR offers, Hopper jotted “and maybe this in uniform,” followed by “they think maybe they can get me a waiver on a promotion.” It seems to reason that the ONR option called for Hopper to return to active duty, a proposition that appeared to be of significant interest for the reserve officer.77