Darwin Among the Machines (19 page)

Flexner believed in generous remuneration for the Institute's administration and permanent faculty, noting that although wealth might invite distractions from a scholar's academic work, “it does not follow that, because riches may harm him, comparative poverty aids him.”
¹²
This generosity was not extended to visitors, “for on high stipends members will be reluctant to leave.”
¹³
Flexner's stipulation that “no lines are drawn between professors, members, or visitors . . . who mingle so freely as to be indistinguishable”
¹⁴
was upheld in the academic sphere, but in salary and housing the distinctions were made clear. Permanent faculty occupy substantial homes scattered among nearby streets and lanes, flanked by investment bankers, pharmaceutical heirs, successful New Jersey gangsters, and others who have managed to pay their dues. Rights of first refusal to these properties are retained by the Institute, extending its influence well beyond the eight hundred acres that make up its own grounds. Institute visitors, in contrast, are billeted in nondescript four-plex apartments, known locally as “the Project,” circumscribing a communal launderette and some worn-out lawns. From a low hill (barely enough for winter sledding) Olden Manor, converted into the director's residence and surrounded by coach house, riding stables, and servants' quarters, overlooks the whole preserve.

The barn was situated below the project at the end of Olden Lane. It was the largest building on Olden Farm until eclipsed by the construction of Fuld Hall in 1939. An imposing, redbrick Georgian edifice, Fuld Hall housed the Institute's administrative offices, library, dining hall, several hallways leading to faculty offices in adjacent wings, and, on the ground floor, a common room presided over by a grandfather clock and leather-bound armchairs, with chessboards arranged near the windows that overlooked a courtyard facing the Institute woods. Fresh newspapers were skewered every morning on a polished wooden rack. Tea and cookies were served on real china daily at exactly three o'clock. Except for the absence of tennis courts (in the basement, table tennis was played underneath the steam ducts in a bare, concrete room), the resemblance to a private club or a European country estate (or perhaps a sanatorium) was pronounced. “We miss informal contact with one another and are about to remedy this defect by the erection of a building provided by the founders, to be called Fuld Hall,”
¹⁵
wrote Flexner in 1939. “Mathematicians will informally lunch, smoke, chat, walk, or play golf with the physicists and the director. Can any possible form of organization give the flexibility, the intimacy, the informality, the stimulus thus attainable?”
¹⁶
My father observed that although most university professors are parasites feeding on the living, Institute professors are saprophytes, feeding on the dead.

Between the barn and the housing project was a low brick building whose Venetian blinds were always drawn. Graced with the appearance of a power substation or a telephone switching exchange, it was constructed suddenly, in 1947, to house the high-speed electronic computer whose circuits had begun to take shape in 1946 in the basement of Fuld Hall. Operational in 1951 and officially dedicated in 1952, the Institute for Advanced Study computer, or IAS machine, was never christened with an acronym of its own, but its far-flung offspring (ILLIAC and ORDVAC at the University of Illinois, JOHNNIAC at the Rand Corporation, MANIAC at Los Alamos National Laboratory, AVIDAC at Argonne, ORACLE at Oak Ridge, BESK in Stockholm, SILLIAC in Australia, BESM in Moscow, and WEIZAC in Israel, among some fifteen immediate siblings) were. After von Neumann's death in 1957 the project was wound down; the machine was turned over to Princeton University in 1958 and then dispersed, in part to the Smithsonian, in 1960. Although we Institute children competed in exploring every inch of the Institute's domain, from the abandoned pig farm at the far corner of the Institute woods to secret passageways in the basement of Fuld Hall, the computer building was out of bounds. In 1994 the building, still known as the
ECP building, was converted to a day-care center. Now a new generation is growing up in the same nursery where, by the flickering glow of twenty-six hundred vacuum tubes, the progenitor of the microprocessor was born.

After conflicts over patent rights divided the members of the ENIAC and EDVAC project at the Moore School, Eckert-Mauchly and von Neumann went separate ways. Von Neumann realized that the most direct way to influence the design of electronic computers, and to secure their unrestricted use, was to construct one for himself. “If he really wanted a computer,” explained Arthur Burks, “the thing to do was to build it.”
¹⁷
Von Neumann structured the project so as to introduce multiple copies of the new machine in several places at once. Progress reports were disseminated not only among the participating funding agencies and to a half-dozen groups that were duplicating the IAS design, but to any location where the potential of high-speed digital computers might fall on fertile ground. It is no accident that the vast majority of computers in circulation today follow the von Neumann architecture—characterized by a central processing unit operating in parallel on the multiple bits of one word of data at a time, a hierarchical memory ranging from fast but limited random-access memory to slow but unlimited media, such as floppy disks or tape, and a distinction between hardware and software that enabled robust computers (and a robust computer industry) to advance by a leapfrog process with each element evolving freely on its own.

At the time of von Neumann's project there were several competing architectures at the gate. Von Neumann chose, and spurred, the winning horse. “He was in the right place at the right time with the right connections with the right idea,” explained Willis Ware, “setting aside the hassle that will probably never be resolved as to whose ideas they really were.”
¹⁸
Although von Neumann did not foresee the personal computer, he did foresee, in agreement with Wiener's
Cybernetics
, that “science, as well as technology, will in the near and in the farther future increasingly turn from problems of intensity, substance, and energy, to problems of structure, organization, information, and control.”
¹⁹

MIT and the University of Chicago both offered substantial incentives to attract von Neumann, who remained determined to build his computer at the Institute for Advanced Study, despite the absence of laboratory facilities of any kind. “The Institute is a place which works with blackboards, paper and pencil, not with physical instruments and experimental techniques,” recalled Willis Ware in 1953. “So the coming of six engineers with their assortment of
oscilloscopes, soldering irons, and shop machinery was something of a shock.”
²⁰
In closing down the computer project after von Neumann's death, the trustees would reemphasize their commitment to pure research and advise that no experiments ever be permitted again. Institute colleagues, said Arthur Burks, “thought that von Neumann was not using his creative mathematical skills properly in being involved in computers.”
²¹
Von Neumann was determined to prove them wrong. There were preliminary discussions with Princeton University concerning laboratory resources and staff support, but little materialized, although RCA's Princeton Laboratories made a firm commitment and did contribute technical support. With Aydelotte on his side, von Neumann was persuasive in his argument that the high-speed computer was a revolutionary instrument that would change the nature of mathematical research. In every field of science, von Neumann could cite specific instances in which a computational focus would give insight into the unknown. He had ready access to outside funding and was going to build a computer, either at the Institute or somewhere else. The trustees did not want to lose von Neumann to another institution, so they gave him the go-ahead.

In November 1945, a committee led by von Neumann and including Herman Goldstine (still attached to the ENIAC project at the Moore School) held its first meeting in Vladimir Zworykin's office at RCA. Also in attendance was John Tukey, a statistician with Bell Telephone Laboratories (and the originator of information theorist Claude Shannon's landmark contraction of “binary digit” to “bit”). Von Neumann issued a memorandum of objectives, concluding that “it is to be expected that the future evolution of high-speed computing will be decisively influenced by the experiences gained.”
²²
By the spring of 1946 the project was under way and staff, led by Goldstine, were signing on.

Goldstine, Arthur Burks, and von Neumann set to work developing the logical plan of the computer, released in June 1946 as
Preliminary Discussion of the Logical Design of an Electronic Computing Instrument
, revised in September 1947, and followed by a three-volume report,
Planning and Coding of Problems for an Electronic Computing Instrument
(1947–1948). These documents were circulated even more widely than the EDVAC report, setting precedents that computer architecture and programming has followed ever since. “The remarkable feature of the reports,” according to I. J. Good, “was that they gave lucid reasons for every design decision, a feature seldom repeated in later works.”
²³

Von Neumann worked out of his office at the Institute, and Goldstine and Burks were installed in an annex to Gödel's office down
the hall. “Kurt Gödel didn't have a secretary, didn't want one, I assume,” recalled Burks. “So for that summer, when of course we didn't yet have a building for the computer, Herman and I occupied the secretary's office next to Gödel's office. It had a blackboard on the wall.”
²⁴
This was more than could be said for the facilities provided the engineers, who were given a bare storage room in the basement of Fuld Hall. “Our first job was to build work tables for us to work on,” recalled Ralph Slutz, a Princeton graduate student who joined the group in June after completing his degree. “We asked von Neumann if he would pay for the paint if we painted the walls a more reasonable color than they were when we moved in. This he did.”
²⁵
Willis Ware, a twenty-eight-year-old electrical engineer who also arrived in June, recalled the “temporary space in the second basement, surrounding the boilers. You know—it wasn't bad since it was summer and they were turned off. And then we moved upstairs to the first basement under one wing.”
²⁶

Von Neumann had no experience—or desire to experience—building anything with his own hands. Whether at a bomb test in the desert or at a meeting of the trustees, he invariably appeared in a three-piece suit, wearing the uniform of a banker among colleagues who formed a generally disheveled troop. “Once he understood the principle of it, the ghastly details like the fact that you'd have to put by-pass condensers on things, and all sorts of dirty engineering things—that didn't really interest him,” noted Goldstine. “He would have made a lousy engineer.”
²⁷
To translate his logical plan into reality he needed the help of those who could design and build not only the logical control, arithmetic unit, and high-speed memory, but everything else from stable power supplies to air-conditioning to some means of reading data in and out of the machine. As chief engineer von Neumann selected Julian Bigelow, thirty-three years old and “a quiet, thorough New Englander” in the opinion of Norbert Wiener, with whom Bigelow had collaborated during the war on reed-time computing for anti-aircraft fire control. Wiener recommended Bigelow for the job. “We telephoned from Princeton to New York, and Bigelow agreed to come down in his car. We waited till the appointed hour and no Bigelow was there. He hadn't come an hour later. Just as we were about to give up hope, we heard the puffing of a very decrepit vehicle. It was on the last possible explosion of a cylinder that he finally turned up with a car that would have died months ago in the hands of anything but so competent an engineer.”
²⁸

Bigelow was a theoretician as well as a mechanic, and a founding member of the cybernetics group. With Norbert Wiener and Arturo
Rosenblueth he coauthored a 1943 paper, “Behavior, Purpose and Teleology,” suggesting unifying principles underlying intelligent behavior among living beings and machines. “A further comparison of living organisms and machines . . . may depend on whether or not there are one or more qualitatively distinct, unique characteristics present in one group and absent in the other,” concluded Bigelow. “Such qualitative differences have not appeared so far.”
²⁹
This paper served as the namesake for the informal Teleological Society out of which the Macy conferences and what came to be known as the cybernetics movement took form. “Cybernetics came into its own,” explained Warren McCulloch in 1961, “when Julian Bigelow pointed out the fact that it was only information concerning the outcome of the previous act that had to return.”
³⁰

Bigelow, who published little, exercised his influence by linking the mathematical worlds of Wiener and von Neumann with the world of practical machines. He had a gift for building things that worked and fixing those that broke. He was perhaps the only permanent member of the Institute who worked on his own car. “I remember one day walking out the back door of that little brick building,” remarked Ware, “and here's Julian lying under this little Austin, welding a hole in a gas tank. And he said, ‘Nope! It won't explode!' And he had some perfectly reasonable explanation for why it wouldn't explode, based on the principles of physics.”
³¹

Bigelow's job was to take the logical design as laid out in the abstract by von Neumann and coax it to life as a machine. “Julian would have the ideas, or Ralph [Slutz] would kind of detail the ideas, and then Pom [James Pomerene] and I would go try and make the electrons do their thing,” said Ware. “He was kind of more physicist and theoretician than engineer. . . . In modern parlance, what you'd say was: Julian was the architect of that machine.”
³²
His responsibilities included designing and fabricating circuits, building test equipment that could track events in the microsecond range, requisitioning scarce electronic parts, and securing something even scarcer in Princeton—living quarters for his staff. Bigelow managed to find some war-surplus housing and had it moved on site, against the protests of Princeton citizens who felt this might bring the wrong kind of architecture to their side of the tracks. Using their small machine shop, Bigelow's crew constructed not only the main frame of the computer but peripheral equipment as well, such as a forty-four-track high-speed magnetic drum (storing 2,048 40-bit words) and a high-speed wire drive that coiled and uncoiled magnetic recording wire at up to one hundred feet (or 90,000 bits) per second from a pair of
bicycle wheels, differentially coupled side by side on a single concentric drive. The wheels could be removed and inserted as a unit, just as a disk cartridge or other removable medium is used today.