When Computers Were Human (47 page)

Among the members of the Applied Mathematics Panel, only Warren Weaver was interested in taking responsibility for the Mathematical Tables Project. Weaver went looking for more support by polling the leaders of the National Defense Research Committee divisions to see if they might make use of a large computing office. On January 14, he sent a mimeographed letter to the nineteen leaders and waited for the replies. In less than two weeks, he had four votes in favor of the Mathematical Tables Project and fifteen votes against. As he reviewed the comments from the different divisions, Weaver decided to ignore the poll and his own concerns about security. “Of my own knowledge,” he told the Applied Mathematics Panel, “I can say that if Dr. Lowan's group were disbanded, another group of several dozen computers (at least) would have to be set up in NDRC somewhere to take care of the calculations which Dr. Lowan is at present carrying on for various sections of the NDRC.”
⁹³

For a time, Weaver went looking for an alternative to the Mathematical Tables Project, a computing office that had no history of work relief and no baggage of security problems, but there was no organization with a similar expertise in scientific computation. “I hear that you have considerable computing machines at Vassar [College] and that you have some experience with them,” he wrote to a mathematics professor named Grace Hopper (1906–1992). “Are you interested in doing work for the Applied Mathematics Panel?”
⁹⁴
Hopper was available for the work, but she was hoping to win a commission with the navy. The navy was resisting her overtures, claiming that she was too old, at the age of thirty-six, and was underweight. Undaunted, Hopper persisted and finally received her commission. The navy assigned her to Howard Aiken's computing facility at Harvard, which handled a variety of calculations for weapons and communications research.
⁹⁵

Unable to find any alternative to the Mathematical Tables Project, Weaver began to bring the group under the control of the Applied Mathematics Panel and was surprised to discover that Arnold Lowan was circulating his own plan for the project. Lowan proposed two options to the Applied Mathematics Panel. The first would establish a group of six
mathematicians and fifty computers; the second would keep the same number of mathematicians but retain only twenty-five computers. When asked about the plan, Weaver confessed that he was “just a little embarrassed by this whole situation.” He had no direction from the National Defense Research Committee, and “my only information concerning the budgetary possibilities is contained in an estimate which I believe Dr. Lowan prepared … and a copy reached me by an indirect route.” After studying the document, Weaver decided to accept the smaller of the two plans. With a single letter, he swept away all the restrictions of the WPA and most of the stigma of work relief. The Mathematical Tables Project would be a contractor to the Applied Mathematics Panel. It would operate as an office of the National Bureau of Standards, though Lowan would take his orders from the panel. The agenda for the group was no longer set by a New Deal agency in Washington but by a committee of well-respected mathematicians.
⁹⁶

Even though Arnold Lowan would have to sacrifice seventy-five computers, he was pleased with the offer from the Applied Mathematics Panel and told Philip Morse that this “very satisfactory arrangement is unquestionably due to a great extent to your constant efforts on our behalf, for which please accept the expression of our warm gratitude.”
⁹⁷
Lowan had about two weeks to finish his obligations to the WPA and prepare for his new assignment. March 15 was the last day of operation as a relief project. The computers packed their equipment, burned old computing sheets, and disposed of the posters that had guided them. As the Mathematical Tables Project shed the seventy-five computers, it also shed the manual computing division. When the project opened its new office, all computers would use adding machines or mechanical calculators.
⁹⁸
Ida Rhodes would identify this move as the time that “life began,”
⁹⁹
the moment when the Mathematical Tables Project finally dropped the trappings of work relief and become a professional computing organization.

CHAPTER SIXTEEN

The Midtown New York Glide Bomb Club

T
HE WINTER OF
1943 marked the start of the imperial age of the human computer, the era of great growth for scientific computing laboratories. It seemed as if all the combatants discovered a need for organized computing that winter. A German group started preparing mathematical tables at the Technische Hochschule in Darmstadt.
¹
Japan, which had received material from the Mathematical Tables Project through 1942, formed a computing group in Tokyo.
²
The British government operated computing groups in Bath, Wynton, Cambridge, and London.
³
Within the United States, there were at least twenty computing organizations at work that winter, including laboratories in Washington, Hampton Roads, Aberdeen, Philadelphia, Providence, Princeton, Pasadena, Ames, Lynn, Los Alamos, Dahlgren, Chicago, Oak Ridge, and New York City. Most of these calculating staffs were small, consisting of five to ten computers. Langley Field, a major aeronautical research center in Virginia, employed about a dozen such groups, each assigned to a specific research division. “Some [groups] have as many as ten computers,” explained a history of the center, “while others have one computer who often devotes a part of her time to typing and secretarial duties.”
⁴
Only a few computing laboratories were as large as the New York Hydrographic Project with its forty-nine veterans of the Mathematical Tables Project or the thirty-person computing office of the Naval Weapons Laboratory at Dahlgren, Virginia.
⁵

Amidst this growth of computing offices, the MTAC committee finally came to life and began to chronicle the literature of calculation. Nearly eighteen months after his confrontation with Luther Eisenhart and the National Research Council forced him to retreat, R. C. Archibald had returned to his post in the summer of 1942 and announced a new goal for the committee. “Our
Guides
are very slow in appearing,” he wrote. “Hence I have been led to the conviction that it would be very desirable to establish a quarterly publication called
Mathematical Tables
” in order
to circulate the committee's bibliographies and reports.
⁶
The proposal surprised the members of the National Research Council. It would “be like issuing a professional journal,” complained the permanent secretary.
⁷
After the argument over the $61.73, the council members were uncomfortable with the idea and tentatively tried to check Archibald. They approached Warren Weaver, in his role with the Rockefeller Foundation, and asked if the money granted to Archibald by the foundation could be used to finance a publication. Weaver confessed that he had a “certain horror” at being associated with Archibald's idea, but he also stated that the Rockefeller Foundation would not stop the new periodical.
⁸
The members of the MTAC committee, from L. J. Comrie to Charlotte Krampe, were slower to respond, but they generally liked the proposal. Wallace Eckert wrote that the periodical “would serve a very useful purpose” but warned that it “would probably become a financial headache” and that the “present is not the most auspicious time to start it.”
⁹

Never one to wait for favorable times, Archibald pressed ahead, leaving even the most sympathetic members of his committee behind. “With the load I have to carry,” he wrote to the MTAC committee, “I can not possibly undertake either to discuss everything with you before hand or send all copy to you before publication.”
¹⁰
He completed the first issue of the journal, entitled
Mathematical Tables and Other Aids to Computation
, in February 1943. To his credit, he recognized that the journal could not flourish if he was the sole contributor and apologized to his readers, “R. C. A. greatly regrets the apparent necessity for numerous personal contributions in this issue, as well as in the second.”
¹¹
The issue contained a great deal of useful information, including lists of tables, errata, book reviews, and articles on methods of calculation. The only thing that seemed out of place was a piece devoted to the computing machines of the seventeenth century, a favorite subject of Archibald's.

Mathematical Tables and Other Aids to Computation
provided American computers with the first systematic reports on computing activities. Before the journal reached a wide audience, many computers did not know what organizations existed and what work was being done. In early February 1943, the members of a new computing group at the University of Pennsylvania did not know how they might contact the Mathematical Tables Project. One of the group's leaders, John Brainerd (1904–1988), sent a letter to the project sponsor, Lyman Briggs at the National Bureau of Standards. Brainerd explained that he was undertaking a large computing effort for the Aberdeen Proving Ground and was searching for human computers and computing expertise. He hoped that the Mathematical Tables Project was still operating and that it might provide him with human computers or handle some of his calculations or provide him with training materials.
¹²
Brainerd needed especially sophisticated computers,
computers with a good background in mathematics. The Mathematical Tables Project might have seemed an unlikely source of such computers, but Gertrude Blanch had initiated an extensive training program in 1941. She and other members of the planning committee developed a series of eight mathematics courses, which they offered over the lunch hour. The first course discussed the properties of elementary arithmetic; the intermediate ones covered standard high school algebra, trigonometry, and college calculus; the final course presented the methods of the planning committee: matrix calculations, the theory of differences, and special functions. The teachers treated the courses as a formal school, requiring the students to attend every session and asking them to “do a reasonable amount of ‘home work' on their own time.”
¹³

Lyman Briggs replied to Brainerd's letter just as the Mathematical Tables Project was preparing to move from its old WPA office to the rooms rented by the Applied Mathematics Panel. He explained to Brainerd that the project had found a home for the duration of the war and was able to accept outside assignments. “I think you will be glad,” Brainerd told his colleagues, “to note the action which is being taken in connection with the computation project.”
¹⁴
The enthusiasm of this initial contact quickly faded as the leaders of the two computing organizations employed different strategies in their work. Arnold Lowan, of the Mathematical Tables Project, was a classical physicist who understood the rules of divided labor. For him, computing machinery was an aid that “facilitated and abridged” the efforts of his staff. Brainerd was a professor of electrical engineering at the University of Pennsylvania. He organized his office around a large computing machine, a differential analyzer, and used human labor to compensate for the machine's shortcomings. Brainerd's computers were machine operators, as George Stibitz had prophesied, but these operators were not mere drudges, for they needed a thorough mathematical education in order to do their work.

In 1937, the University of Pennsylvania had acquired a differential analyzer in conjunction with the Aberdeen Proving Ground. The proving ground had financed the differential analyzer under an agreement that allowed ballistics researchers to use this machine in times of war. Until the spring of 1942, the analyzer had been used by engineering professors and graduate students. Like most university research equipment, this machine received regular but intermittent use. Four or five times a term, it would calculate a curve associated with some electrical component or circuit. Occasionally, it would serve as the object of an experiment by a graduate student interested in electromechanical controls. Once or twice a year, the university was able to rent the device to a local company. For other periods, the machine stood idle, gathering dust and dripping oil.
¹⁵

In June 1942, proving ground officials notified the University of Pennsylvania
that they needed to use the differential analyzer for ballistics research and offered to reimburse the school $3.00 an hour for operational costs: electricity, the wages of mechanics, supplies, and the salaries of any staff that were needed to oversee the calculations. A small group of Aberdeen researchers took the train north from the proving ground to inspect the machine. The analyzer was housed in a nondescript brick building just a few blocks from the railroad station. Their first test of the machine, a trajectory for 4.7” antiaircraft shells, was disappointing. “Upon arrival,” wrote a member of the Aberdeen staff, “it was apparent that a desirable rate of analyzer output had not been achieved.” The output from the machine substantially deviated from a hand-calculated trajectory. “The Philadelphia analyzer … has not been under the compulsion of the great accuracy demanded at Aberdeen,” observed a proving ground researcher, “and therefore has not been as assiduously cared for as the Aberdeen analyzer.” To “attempt to maintain [high accuracy] with the Philadelphia analyzer,” concluded the army, “required an exorbitantly high number of adjustments and test runs.”
¹⁶

At a hastily called meeting between university officials and army officers, John Brainerd presented a plan that would produce results within 0.5 percent of hand-computed values. This plan called for a few modifications to the machine, strict operational standards, and a staff of human computers to oversee every step of the calculation.
¹⁷
An early test of the new procedure achieved the specified accuracy but at the cost of substantial hand calculation. It is “desirable to expand the Philadelphia unit somewhat at once,” concluded the army, in order “to train and prepare its personnel for handling the contemplated output of the analyzer.” The calm words of the military report camouflaged the problem facing the Pennsylvania faculty. The university did not have enough college-educated computers for its analyzer staff. They had hoped to find twenty to thirty women with bachelor's degrees in mathematics or physics, but after scouring the school's alumna lists, they had identified only eight who held the appropriate degree. Brainerd had offered each of them a position as an assistant computer with a salary of $1,620 per annum, but he believed that no more than three or four would accept these positions. With no other obvious options, Brainerd concluded that the university would have to prepare a curriculum for human computers and operate training classes.
¹⁸
He found money for this endeavor at the government's Engineering, Science, and Management War Training Program and borrowed course materials from Aberdeen veteran Gilbert Bliss, who had taught ballistics classes to civilians at the University of Chicago.
¹⁹