Chapter Five · Part II, The Builders

The Man from Budapest

Published

Herman Goldstine, an army lieutenant waiting for a train at Aberdeen, Maryland, recognises the most famous mathematician in the world on the platform and mentions a machine in Philadelphia. John von Neumann's visit became a 101-page draft report with one name on it, a priority war with Presper Eckert and John Mauchly, and the design every computer has used since: instructions kept in the same memory as the numbers. He ended his life under military guard, working on how a brain differs from his own machine.

"The whole atmosphere of our conversation changed from one of relaxed good humour to one more like the oral examination for the doctor's degree in mathematics."

Herman Goldstine on meeting John von Neumann at Aberdeen station, 1944

In the early summer of 1944, on the open-air platform of the railway station at Aberdeen, Maryland, a thirty-one-year-old American army lieutenant named Herman Goldstine stood waiting for a train back to Philadelphia and recognised, on the same platform, the most famous mathematician in the world.

The mathematician was forty years old, short, plump, dressed in a three-piece suit despite the heat, and carrying a briefcase. He was waiting for a train to Los Alamos. Goldstine knew his face from photographs and from the corridors of the Aberdeen Proving Ground, where the man had been a consultant on artillery and explosives for two years; they had never spoken. Goldstine, by his own later account, was nervous.

He approached. He introduced himself. The mathematician, who had a gift for making strangers feel at ease, made small talk about the heat and the war and the shortage of decent coffee. Goldstine, encouraged, mentioned what he was working on in Philadelphia. He was the army liaison to a project at the University of Pennsylvania to build an electronic computer that would, when finished, perform three hundred and thirty-three multiplications a second.

The mathematician's face changed. The small talk stopped. He began to ask Goldstine questions. After three minutes Goldstine wrote, decades later, that "the whole atmosphere of our conversation changed from one of relaxed good humour to one more like the oral examination for the doctor's degree in mathematics." The two trains came and went. The mathematician asked when he might come to Philadelphia and see the machine. Goldstine, dazed, said any time, any time at all.

The mathematician was John von Neumann. The visit he proposed on the platform that day became, within fifteen months, the redesign of computing.


He had been born in Budapest on 28 December 1903, on Vaci utca in the centre of Pest, the eldest son of a banking lawyer named Max Neumann who had three years earlier taken his children's tutoring out of the German governesses' hands and given it to a private team. Max bought the family a hereditary noble predicate in 1913, the margittai of margittai Neumann János, the von of the German rendering, for what was, even by the standards of Habsburg honours-trafficking, a serious price. He could afford it. The Kann-Neumann household, his wife Margaret's family included, occupied the entire top floor of an apartment block above the city, with three brothers, two sets of grandparents, four governesses and a kitchen staff. The boy who would become John was called Jancsi.

He was the kind of child who is occasionally born. At six he could divide eight-digit numbers in his head and joke with his father in classical Greek. He had a memory which his closest friends, in adulthood, would describe as not quite human: he could be handed an opened book and asked to continue reading aloud at any line, and he would, and then if asked he would translate the next paragraph into German or French or Latin or Hungarian without pausing, and he could do this with books he had read decades earlier. The party trick was A Tale of Two Cities, which he had memorised at twelve; he was still doing it at fifty.

In this chapter
John von Neumann, Herman Goldstine, J. Presper Eckert, John Mauchly, Klári von Neumann, Julian Bigelow, Arthur Burks, Stanislaw Ulam
The idea it builds
The stored program: memory as a row of numbered pigeonholes holding instructions and numbers alike, so that a program can write a program.