Astin conflicted with Secretary Weeks over an event that was interpreted one way by National Bureau of Standards scientists and quite another way by members of the business community. It concerned a report by bureau scientists on battery additives, chemical mixtures that promised to enhance the performance of automobile batteries. The report reviewed one such additive, called AD-X2, and judged that it had no value.50 To the scientists, the report was a simple scientific conclusion. To the manufacturer of AD-X2, the report was unnecessary government interference in the marketplace. The manufacturer found enough sympathetic ears in Congress to provoke a debate on the report and put the National Bureau of Standards on the defensive. Allen Astin supported his scientists, much to the dismay of his boss, Secretary Weeks. Concluding that Astin was just one more example of “deadwood and poison oak,”51 Weeks asked for his resignation.
Weeks did not anticipate the extent to which the scientific community would fight for its autonomy, its ability to make decisions without external interference. He also did not foresee the scale of political connections that the scientists could rally. The key players were not the scientists in the National Bureau of Standards, who might be expected to resign in protest, but outside researchers with an interest in the agency. The fight was led by the board of visitors, the scientists and industrialists who offered their advice on bureau operations. “The most influential member of the committee, who seemed to have Mr. Weeks’ respect, was Mervyn Kelly, then President of Bell Telephone Laboratories,” recalled Astin.52 They also drew support from university presidents, academic scientists, and members of the press, such as Washington Post columnist Drew Pearson.53
Faced with a storm of bad publicity, Weeks asked Astin to return “on a temporary basis.”54 Astin used the opportunity to strengthen the hand of bureau scientists. He asked for a full review of the bureau by the National Academy of Sciences. The report, released in the summer of 1953, validated the position of the National Bureau of Standards in the AD-X2 matter, but it concluded that the bureau faced a fundamental conflict between its military research and its civilian duties. It recommended that the bureau return all military-sponsored research to the military agencies. Astin accepted the report and implemented its recommendations. He purged the bureau of all military research, one-third of its budget. The Institute for Numerical Analysis was one of the first units touched by the order. Though the institute appeared to be a civilian laboratory located on a university campus, it was actually financed by money from the air force and the navy. Astin announced that the institute would be closed in June 1954, that its equipment would be given to UCLA, and that its computing staff would be dispersed.55
Anticipating the demise of the Institute for Numerical Analysis, Gertrude Blanch resigned her position at the end of 1953. The computing office at the institute had started to shrink even before Astin announced the closure of the UCLA office. The engineers and researchers who needed computing services had begun acquiring calculating equipment of their own. The most common computing machines were the IBM 604 multiplying punch and its near relative, the Card-Programmed Calculator. The larger customers had ordered IBM’s first electronic computer, the Model 701, which had arrived on the market a year before. Fully six of the first twelve customers for the 701 were companies that had once requested computations from the Institute for Numerical Analysis.56
Blanch could have sought a position with IBM, a job that probably would have taken her back to New York City, but instead she accepted an offer with the mathematics department of the ElectroData Company. ElectroData was located in nearby Pasadena and proclaimed that “mathematicians are the heroes of the new industrial revolution.”57 The firm had been founded by Herbert Hoover and had built electronic instruments during the Second World War. In 1954, it was trying to enter the computer business by building a small machine for engineering applications, the kind of work that was done by the firms that had once patronized the Institute for Numerical Analysis. The company had assembled a talented group of scientists, though none of them belonged to the inner circle of computer designers, the group that could trace their knowledge of computing machines back to the Moore School Lectures of 1946. In addition to Blanch, the office included Clifford Berry, an engineer who had helped John Atanasoff build his computing machine at Iowa State College, and Ted Glaser, a young physicist who was proving to be an exceptional computer scientist even though he had lost his sight as a child.58
The winter of 1954 was a poor time to be a government scientist and an equally inauspicious moment to join a small and inexperienced computer company. The managers of ElectroData were struggling to control the design of the machine, to keep the project on schedule, and to develop a customer base. Shortly after Blanch arrived at the company, the senior managers realized that their company needed computer programmers more than it needed mathematicians, and they replaced the “Mathematics Department” with the “Technical Services Group.”59 After a couple of weeks on the job, Blanch concluded that “the place in Pasadena wouldn’t last too long,” but she did not know where she might find another job.60 Most of the old computing groups had been closed, and their leaders had retired. The dean of human computers, L. J. Comrie, had died. R. C. Archibald had resigned from Mathematical Tables and Other Aids to Computation.61 H. T. Davis was preparing a final volume of mathematics tables, but his numbers had been calculated by his National Youth Administration computers and did not represent contemporary work. Clara Froelich had left Bell Telephone Laboratories, said good-bye to calculation, and departed for an extended vacation in Mexico.62
For Blanch, one of the few benefits of working for ElectroData was the contact it gave her with air force scientists. The company was a contractor for the air force and worked with many scientists who had known Blanch at the Institute for Numerical Analysis or the Mathematical Tables Project. One of these scientists was Knox Milsaps, the chief mathematician of the air force and a founder of the new Aerospace Research Laboratories at Wright Field in Ohio. During the war, Milsaps had worked with Blanch on several mathematical problems, notably one with an expression called the Mathieu function. In the winter of 1954, he was a regular visitor to the ElectroData offices, as he was planning on purchasing one of the new machines for the aerospace laboratories. As he passed through the building, he would often stop at Blanch’s desk, chat about old times, ask about her new work, and inquire, “When are you coming to Wright Field?” At first, she deflected his questions. She was happy in California, and after her appearance before the Department of Commerce Loyalty Board, she was reluctant to expose herself to another security examination. Yet Milsaps was persistent, ElectroData was in financial turmoil, and Blanch recognized that her job was not interesting. “So one night I faced the floor,” she recalled, “and decided to accept the job in Wright Field.”63
While Blanch prepared to move to Ohio, the FBI began one more review of her background. The agents in charge of her case found the results of the earlier investigations and decided that they should reopen each of the charges. In addition to the five issues that had been dismissed by the Loyalty Board, they requested a study of one more concern: “Subject is not known to be married or have ever been married.”64 By this time, the political wars of the early 1950s were starting to wane. The Republicans were more comfortable with power, and the most violent of the anticommunist voices had lost some of their control over the public and their party. The Washington FBI office rejected the request, stating that a new investigation would have a “greater possibility of embarrassment to the Bureau … than [of] attaining any information of value from the subject.”65 They formally closed her file and granted Blanch the full security clearance needed for her work. She moved to Dayton, became head of mathematical research at the laboratory, and “learned to drive on icy streets the same as everybody else.”66
Before Gertrude Blanch could settle in Dayton, before she could buy a house and learn the neighborhoods, she was contacted by Phil Morse about a project that he c
alled the “handbook for the ordinary computer.”67 Morse had concluded that the new electronic computers would be of little help to his generation of scientists and engineers. He speculated that a decade would pass before the typical scientist would have easy access to them. From observing the machines at MIT, he had learned that “programming took weeks, not minutes,” and that small and midsized problems could be handled much more quickly by hand.68 Morse told Blanch that he wanted to hold a conference to discuss such computing problems and asked who should be invited to it. Blanch gave him a long list of “tablemakers,” as she described the computers, who worked at the Aberdeen Proving Ground, the navy’s proving ground at Dahlgren, Virginia, the Atomic Energy Commission, and the University of Illinois, though she was quick to characterize one Illinois faculty member as “anti-table.”69 She did not list Arnold Lowan, as she was not quite sure how he would feel about the invitation. Lowan had already received an invitation from Morse and had replied that “my chief interest in the conference in mathematical tables derives from the assumption that perhaps it is still possible to bring to life the ‘Math Tables Project’ in New York.”70 After Morse explained that he had no interest in reviving the Mathematical Tables Project, Lowan declined to come.
When the conference met in September 1954, most of those in attendance were senior computers of the Second World War: Gertrude Blanch, John Todd, Milton Abramowitz, and even Blanch’s former critic Wallace J. Eckert, who had become an employee of the IBM Corporation. The list also included Nicholas Metropolis from the Manhattan Project; John Tukey, one of the Applied Mathematics Panel mathematicians from Princeton; and the recently fired John Curtiss, who had a temporary job at New York University. Adele Goldstine, who had worked with the computers at the University of Pennsylvania, came only as a spouse. Morse had invited her husband, Herman, because of his contributions to the ENIAC.71
As they discussed the problems of midsized calculations, the conference members quickly agreed that they needed to produce a book containing “tables of usually encountered functions” as well as graphs, mathematical analyses, and “other techniques useful to the occasional computer.”72 They gave the project a shorthand name, the “new Jahnke-Emde,” a term which referred to the book Funktionentafeln mit Formeln und Kurven (Tables of Functions with Formulae and Curves), by Eugen Jahnke and Fritz Emde.73 This book was nearly fifty years old but remained popular with those who worked with applied mathematics.74 Twice during the 1940s, the National Bureau of Standards had suggested revising the book. The first time, in 1941, the bureau was preparing for war and could not find the money for a revision. The second time, in 1947, bureau scientists were organizing their new applied mathematics laboratory and were confidently predicting that the electronic computer would remove the need for such a book. Seven years later, when questioned by Phil Morse, the leaders of the bureau acknowledged that the electronic computers were not the solution to all calculating problems and that they could find some money for a new Jahnke and Emde.75 The book would be prepared by the bureau’s Computation Laboratory and would be edited by Milton Abramowitz and Irene Stegun, two of Blanch’s closest lieutenants from the Mathematical Tables Project. It would contain twenty-two chapters, each written by a different volunteer. As the conference came to a close, Blanch agreed to write a chapter on Mathieu functions. After a little prodding, the sulking Arnold Lowan agreed to contribute a chapter as well.76
The project was formally named the Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, but the mathematicians who worked on the book usually called it “the handbook” or “AMS 55,” after its place in the National Bureau of Standards publication list, or “Abramowitz and Stegun,” after its editors. Milton Abramowitz did the bulk of the preliminary work and recruited most of the contributors, but he died on a hot and sticky July day in 1957, when he imprudently attempted to mow his lawn in suburban Washington, D.C.77 Irene Stegun retrieved the plans for the book from her partner’s desk and finished the project. She corresponded with the contributors, corrected their chapters, and slowly pulled the material into a complete reference work.
The handbook required a decade to complete, a decade that marked a radical change in the electronic computer. In 1954, computers were handcrafted devices that could be found only in government laboratories and large businesses. By 1964, the year of the book’s publication, computers were standard products that could be purchased from a dozen different vendors. The actual date of publication coincided with the announcement of the IBM System 360, a family of machines that IBM chairman Thomas J. Watson Jr. proudly called “the most significant product announcement in IBM history.”78 The System 360 would anchor IBM’s product line for twenty-five years and would move computing into offices and laboratories that had never had access to the machines of 1954.79 To those who were promoting the new electronic computers, the Handbook of Mathematical Functions seemed an anachronism, a tool for modern science that had been produced by the old human computers. Scientific calculation had become a small part of the subject known as computer science. Computer scientists were increasingly interested in databases, sorting methods, the manipulation of text, and the representation of human reasoning. The tables and formulas of the handbook appealed only to a small group of computer researchers.
Though it no longer represented the central issues in computation, the Handbook of Mathematical Functions ultimately validated the vision of Phil Morse. Neither the IBM 360 nor any of the other machines announced in 1964 was able to handle all of the small and midsized scientific computations that were found at universities and government research labs. Scientists, still having to do some calculations by hand, turned to the contents of the handbook for assistance. They gave good reviews to the book and purchased copy after copy for their laboratories. Within a few years, it became the most widely circulated scientific text ever published.80 “I don’t know what I’d do without it,” wrote Blanch, who called it “one of the books I would keep, if I go anywhere, even if I don’t look at it again.”81
The Handbook sat on Blanch’s desk until she left government service in 1967. It stayed with her in retirement as she wrote a textbook on calculation. In spite of her prediction, it was not taken on her tours of Europe, though one can easily imagine her discussing references to the book with her traveling companion, Ida Rhodes. It was found in her personal effects after her death. As she looked back on her career, the Handbook of Mathematical Functions was the symbol of all that she had accomplished. It appeared at the perihelion of her time in government service, her moment at the center of power. She had just finished a successful decade with the Aerospace Research Laboratories. She had published twenty mathematical papers, been elected a fellow of the American Association for the Advancement of Science, been promoted to the highest rank possible for a government scientist, and been publicly recognized for her contributions to the air force.82 Blanch had even learned a little about electronic computers, though she was never interested in programming. Programming could be left to assistants. As a newspaper article described her, Blanch was “the brain behind the mechanical brain,”83 but a photo taken at the same time shows her glowering at a new electronic computer.
44. Gertrude Blanch and the electronic computer
45. Gertrude Blanch and President Lyndon Johnson
On March 3, 1964, just a few weeks before the Handbook was released to the public, Gertrude Blanch arrived at the White House in Washington, D.C., wearing the finest of her clothes and the tallest of shoes. Along with five other women, she was ushered through the east entrance, escorted down a long hallway, and asked to wait in a small public room with a presidential photographer and a few representatives of the press. President Lyndon Johnson soon walked into the room, greeted the women, and took his place at a podium. “I believe a woman’s place is not only in the home,” he began, “but also in the House and Senate.”84 Blanch and the other five women glanced at each other and smiled knowingly. None of them held an
elective office, but all of them served in government. They had come to the White House as part of Lyndon Johnson’s tribute to Eleanor Roosevelt, who had died the prior November. Johnson had wanted to emphasize the impact of the former First Lady by honoring six distinguished women in government service who had begun their careers when Roosevelt lived in the White House.85 After the speech, Johnson moved around the room and shook the hands of his honored guests. He paused briefly with Gertrude Blanch, bending over to share a word or two. Like Blanch, he had started his professional career at a work relief agency. Between 1935 and 1937, he had served as director for the Texas office of the National Youth Administration and had distributed research funds to the universities of his state. Had there been more time for president and mathematician to speak, they might have found that they shared much in common, but Johnson quickly strode out of the room and vanished down the hall that led to the Oval Office.
From the White House, a car carried Blanch to a Washington air force base, where a dinner had been prepared in her honor. There, amidst friends and colleagues, she relaxed and smiled and danced. These were rare expressions of emotion for one who kept herself under tight discipline. That night, she enjoyed herself as if the whole world moved in orbit around her, for she had heard her president commend her as the top mathematician in the air force, as a founder of the scientific discipline called numerical analysis, as a patriotic citizen who had served in time of war for the Applied Mathematics Panel, as Gertrude Blanch, who had once worked for the WPA and had once managed a staff of human computers.
When Computers Were Human Page 41
