When Computers Were Human (56 page)

43. John Todd, Olga Taussky, and John Curtiss on excursion from Institute for Numerical Analysis

Most members of the institute staff believed that the human computers would be temporary workers, quickly replaced by a new “automatic computing machine,” as electronic computers were then called. Initially, the National Bureau of Standards intended to purchase a computing machine for the Institute for Numerical Analysis. Three American companies were offering to build such machines: Raytheon Corporation of Massachusetts, Electronic Research Associates of Minnesota, and UNIVAC Corporation of Pennsylvania, which had been formed by the ENIAC designers, J. Presper Eckert and John Mauchly. John Curtiss was
not satisfied with any of these proposals and convinced the senior staff at the bureau that the Institute for Numerical Analysis should build its own machine. He argued that a machine could be built quickly and that it would develop improved computing technologies.
¹⁶

Under the best of circumstances, a speedy construction would take two or even three years after the first designers arrived at UCLA. To provide an interim machine-computing service, the institute acquired a device called the Card-Programmed Calculator. The Card-Programmed Calculator stood halfway between the IBM punched card tabulators and the new electronic computers. It had been created by engineers at an IBM customer, Northrop Aircraft Company. The engineers had recognized that they could create a fairly powerful computing device by connecting a new IBM card tabulator to one of the company's accounting machines. IBM had not intended for the machines to be joined in this fashion, but they quickly realized the value of the combination and adopted it as an official product. They eventually leased about 700 of these machines, far more than any of their early electronic computers.
¹⁷

By the fall of 1948, the institute was offering computing services with the Card-Programmed Calculator to both academic researchers and commercial firms in California, including aircraft manufacturers and oil companies. Within six months, Albert Cahn could report that “calls upon this service are such that the facility has been expanded to almost double the size contemplated when the institute was established.”
¹⁸
Originally, the new calculating machine was overseen by one of Blanch's computers, a woman named Roselyn Seidel, but with the increased demands, the leaders of the National Bureau of Standards wanted the device to be managed by someone experienced with punched card machines.
¹⁹
After brief negotiations, John Curtiss convinced Everett Yowell to leave Columbia University and join the Institute for Numerical Analysis. The mathematical staff considered Yowell a “significant appointment” to the institute, but Yowell recalled that the human computers were less enthusiastic. “Ros [Seidel] was not happy to have me hired over her.”
²⁰

The Card-Programmed Calculator was not a full electronic computer, though it was a considerable improvement over the old punched card tabulators. Using special punched cards and an IBM plugboard, an operator could instruct the machine to undertake a complicated series of operations. It also had an electronic memory that could store forty-eight numbers of ten digits each.
²¹
With this memory, the machine could handle small simultaneous equation problems, such as the least squares calculations that had stymied George Snedecor at Iowa State University twenty years before. However, the new machine could not handle large problems, including the one that had appeared in the Mathematical Tables Project test of linear programming.

Some simultaneous equation calculations posed unusual problems for Blanch's computers. On these calculations, the computers could follow every step of the computing plan, check the work with a desk calculator to ensure than every step was done properly, and still produce values that were wildly incorrect. Though some blamed the computing plan, Blanch discovered a difficulty that would eventually be called “ill conditioning.” Ill-conditioned simultaneous equation problems are fundamentally unstable, just as a coin balanced on its edge is unstable. Rounding the values of an ill-conditioned problem, a simple and innocuous act, can cause the calculation to collapse into a meaningless mess of figures. The only way to fix this problem is to reorganize the computing plan, producing a plan that is algebraically equivalent to the original calculation but avoids certain combinations of the four arithmetic operations.
²²

During the construction of the new machine, the mathematicians at the institute started exploring computing techniques that were difficult to do by hand. They experimented with linear programming and invited John von Neumann to visit the institute. Another mathematician developed “relaxation techniques,” computing methods that began with a rough guess of the final answer and then slowly adjusted that guess. Many at the institute became interested in “Monte Carlo” techniques, methods that used random numbers to calculate answers.
²³
The institute computers worked on such calculations as best they could, though they could rarely handle a large problem. The staff never developed the kind of computing skill that was found at the Mathematical Tables Project, as few computers stayed at the institute for more than a year. The rapid departure of human computers never seemed to bother the mathematicians, as they were looking ahead to the new computing machine.
²⁴

The electronic computer took three years to complete and went through three changes of name. Employees at the institute called the machine “Zephyr,” after the great western wind, while the staff of the National Bureau of Standards in Washington referred to it as “Sirocco,” the hot air from the desert. Eventually, the two groups settled on the government acronym SWAC, which stood for Standards Western Automatic Computer.
²⁵
Though the members of the institute occasionally felt isolated from the academic centers of the East Coast, they generally had no regrets about their distance from Washington. John Todd wrote that a “certain distance from Washington was certainly desirable, for some mathematicians are uncomfortable with strict dress code and regular hours.”
²⁶
Their location may have spared them the need to arrive at work at 8:00 in the morning or to wear formal business clothing, but it did not insulate them from the political turmoil of the late 1940s.

The political conflict of this era was rooted in the looming problems of the Cold War, the legacy of Franklin Roosevelt, and the frustrations of
the Republican Party. The Cold War placed the United States in a contest for global dominance with a frightening and powerful enemy. Roosevelt had reordered the political landscape in a way that kept the Democratic Party in power for over sixteen years. The Republicans, frustrated by the Democratic hold on power, were “traumatized and bitterly divided,” observed the journalist David Halberstam.
²⁷
The Republicans had always been a minority party and had held power only by pulling support away from the Democrats. During the period of Republican dominance in the nineteenth century, conservative leaders had often found it useful to discredit Democratic opponents by calling them secessionists, politicians sympathetic to the old Confederate states. Seventy-five years later, a new generation of Republican leaders accused the Democrats of being communists, agents of the Bolshevik revolution, traitors. The fact that some of the Democrats had actually been communists, or at least had been sympathetic to the Soviet Union, bolstered such charges. By the fall of 1949, Republicans equated communism with treason and pointed to the Soviet atomic bomb program for their proof. The Soviet Union had detonated a nuclear weapon on August 29, 1949, an event that surprised American weapons experts.
²⁸
Only a few weeks before, the Central Intelligence Agency had predicted “that [a Soviet] atomic bomb cannot be completed before mid-1951.”
²⁹

The center of the American political maelstrom was the House of Representatives Un-American Activities Committee, often called HUAC. It was filled with disaffected Republicans who were either valiantly defending the United States from an insidious enemy or trying to make a name for themselves by pinning the label of communism on public figures, depending upon one's point of view. The committee investigated the senior members of the Truman administration, the writers, directors, and actors of Hollywood, and the scientists in government service. In 1947, the House Un-American Activities Committee had begun an investigation of Edward Condon, the director of the Bureau of Standards, an investigation that struck uncomfortably close to the Institute for Numerical Analysis.

Condon had liberal inclinations, though he had never been a member of the Communist Party, and he had worked on the atomic bomb during the war. He spent most of the war at the University of California, but he had briefly served at the Los Alamos laboratory. At Los Alamos, he been engaged in “several arguments about security regulations,” according to historian Jessica Wang, and soon resigned his position. In his resignation, he stated his belief that military control of scientific information was impeding work on the bomb. “To his mind, intellectual freedom and international cooperation were intimately linked,” wrote Wang. “Scientific progess required open communications, free from military requirements
of secrecy.”
³⁰
When the House Un-American Activities Committee learned of Condon's record at Los Alamos and his support for the open dissemination of atomic research, the committee reviewed his case and declared that Condon was “one of the weakest links in our atomic security.”
³¹

In general, the House of Representatives investigated only public figures and senior members of the administration, such as Condon. Junior employees, such as those who worked for the Institute for Numerical Analysis, were investigated by administrative committees that had been established by the Truman administration as a way of deflecting Republican criticism. The committee that oversaw the National Bureau of Standards and the Institute for Numerical Analysis was Department of Commerce Loyalty Board Number Two. In December, this board announced that it would require “pre-appointment loyalty checks of all research associates and guest workers who are located at National Bureau of Standards for more than one week.”
³²
This order included everyone working at the Institute for Numerical Analysis, the administrative staff, the eight permanent researchers, and the twenty annual visitors.

The first member of the Institute for Numerical Analysis to be called before Department of Commerce Loyalty Board Number Two was the senior administrator, Albert Cahn. No record has been found of Cahn's case, and all we can do is speculate that Cahn belonged to the broad class of liberal scientists, a group that worried many loyalty investigators. One of the few concrete facts we know of Cahn is that he signed the July 17 Petition, a document drafted by Manhattan Project scientists after the first test of the plutonium bomb. The petition requested that President Truman delay the use of the new weapon.
³³
Manhattan Project officials annotated one copy of the petition, indicating whether each signatory was important or unimportant to the project.
³⁴
While this act may have brought Cahn to the Loyalty Board, it may also have been only one of several pieces of evidence against him. In all, seventy scientists signed the petition, and some of them were never questioned by an investigative panel.
³⁵

After Cahn was summoned before the Loyalty Board, fourteen months passed before his case was decided. The delay did not bode well for him. During this period, Klaus Fuchs confessed to passing atomic secrets to the Soviet Union; Alger Hiss was convicted of perjury; Mao Tse-tung and his followers established the People's Republic of China; Senator Joseph McCarthy gave a rambling but inflammatory speech in which he claimed to have evidence of communists in the State Department; North Korean troops invaded South Korea; and, finally, the FBI arrested Julius and Ethel Rosenberg on charges of spying for the Soviets. When the Loyalty Board finally reviewed Cahn's case, they concluded that the evidence against Cahn seemed to fit into a broader pattern of threats against the
United States. They judged that the institute administrator was a security risk and placed him on administrative leave.
³⁶

Following Albert Cahn came Gertrude Blanch. Blanch was especially vulnerable, as she had already failed not one but two security investigations. The first investigation had been conducted in 1942, when the Mathematical Tables Project was certified as an essential relief project. That review had denied her a security clearance. The second investigation had occurred in 1946, when Blanch had been invited to join the computing staff at Los Alamos. In that review, the FBI quickly uncovered the results of the 1942 investigation and declared Blanch untrustworthy. When the administrators of Los Alamos received this verdict, they quietly withdrew their invitation.
³⁷
Both of these judgments were part of the record presented to Department of Commerce Loyalty Board Number Two in the spring of 1951. When the board considered her case, they concurred with the earlier decisions and judged her untrustworthy, but before she could be placed on administrative leave, Blanch appealed the ruling and requested a formal hearing before the board. The board accepted her request and scheduled her hearing for May 1952.
³⁸

“There are only a few times,” wrote the 1950s sociologist William Whyte, when an individual “can wrench his destiny into his own hands—and if he does not fight then, he will make a surrender that will later mock him.”
³⁹
By nature, Gertrude Blanch avoided public confrontations. Her moments of strength were private moments. She had quietly postponed a college education in order to support her mother. She had gently guided the poverty-stricken computers of the Mathematical Tables Project. On at least one occasion, she had stood firm against the bluster of Arnold Lowan. In this last situation, Blanch had to confront the charges against her or surrender her place as a scientist. The case against her was based on the five points that had been identified in 1942. The first three were circumstantial. First, the FBI had an informant who claimed to have seen her purchase a copy of the
Daily Worker
sometime during the late 1930s. Next, at approximately the same time, the New York office of the WPA identified her as a “Red.” Finally, she had shared an apartment with her sister and brother-in-law, who were open members of the Communist Party. The last two points were harder to dismiss. During the 1930s, Blanch had registered as a member of the Democratic Labor Party, an organization that the FBI claimed was “captured by the Communists.” She had also signed petitions for political candidates who openly identified themselves as communists.
⁴⁰