The third American calculator project began, appropriately enough for an invention conceived in academia, with a memo. Written in August 1942 by John W. Mauchly, a thirty-five-year-old assistant professor at the Moore School, it was entitled “The Use of High Speed Vacuum Tube Devices for Calculating.” A physicist turned engineer, Mauchly (Mawk-ly) proposed the construction of an “electronic calculator” or “electronic computor” (his spelling) consisting of vacuum tubes pulsing at the rate of at least 100,000 beats a second. Estimating the machine’s computational speed at 1,000 multiplications a second, he explained that it would be able to compute a single trajectory in 100 seconds. Such a calculator would not only outrun the vaunted differential analyzer but, Mauchly wrote, also would be much easier to use and considerably more accurate.
Today, Mauchly’s memo, which was submitted to the Ordnance Department and to the Moore School, makes for rather odd reading. Considering that it resulted in the development of the first general-purpose electronic digital program-controlled calculator, it is surprisingly poorly written and organized, and conceptually unsophisticated. The bulk of the document is devoted to a discussion of the potential speeds of electronic calculation. (Ironically, the figures Mauchly supplies – 1,000 multiplications per second, for example – turned out to be three times too high.) Yet, although he was not the first American to conceive of a calculator built out of tubes (we’ll examine this point later in the chapter), Mauchly was one of the first to realize that an electronic calculator would be capable of tens of thousands of operations a second. While many other crucial matters, such as the device’s internal structure, weren’t even mentioned, it wasn’t because Mauchly didn’t have any ideas on the matter. Rather, it was because he wasn’t thinking of a computer but of an old-fashioned calculator that happened to be made of tubes:
As already stated, the electronic computor utilizes the principle of counting to achieve its results. It is then in every sense the electrical analogue of the mechanical adding, multiplying and dividing machines which are now manufactured for ordinary arithmetic purposes…. It is intended that this analogy shall be interpreted rather completely. In particular, just as the ordinary computing machine utilizes the decimal system in performing its calculations, so does the electronic device.
At the time Mauchly submitted his memo, the production of firing tables was still keeping pace with the output of new guns and an electronic calculator seemed unnecessary. As a result Mauchly’s proposal was ignored. But the situation changed about half a year later, when the BRL began falling behind. In March 1943, Herman H. Goldstine, the lieutenant in charge of the BRL’s computing substation at the Moore School, learned of Mauchly’s ideas during a casual conversation with one of the mechanics who watched over the school’s analyzer. Goldstine went to see Mauchly; Mauchly couldn’t find a copy of his memo but his secretary managed to reconstruct one from her notes. Goldstine, who had been an assistant professor of mathematics at the University of Michigan before the war, immediately grasped the memo’s significance and asked the Moore School to submit a thorough proposal for an electronic calculator.
In response, John Brainerd, working with Mauchly and J. Presper Eckert, Jr., a brilliant young electronics engineer who had often discussed the feasibility of an electronic calculator with Mauchly, wrote a proposal and handed it in to the BRL on 2 April 1943. Negotiations moved very quickly. Within six weeks, the Moore School and the BRL reached an oral agreement on the terms of a contract and, on 5 June, the two parties signed on the dotted line. The Moore School would receive $61,700 immediately and another $88,300 after certain technical hurdles had been passed. Brainerd was appointed project supervisor; Eckert the chief engineer; Mauchly, who continued to carry a full load of classes, the principal consultant; and Goldstine the BRL’s technical liaison.