John von Neumann

By Emre Yıldırım

In the summer of 1945, John von Neumann wrote a 101-page document in a few weeks that became the blueprint for every computer on earth. The 'First Draft of a Report on the EDVAC' described a machine in which both the program and the data it operated on lived in the same memory, in the same binary language, so that a computer could treat its own instructions as data, modify them, and loop. The stored-program concept seems inevitable now; at the time, machines like ENIAC were wired by hand for each new calculation, and the idea that software could be loaded like cargo into a uniform memory was a conceptual detonation. Von Neumann, born János Neumann in Budapest in 1903 to a wealthy banking family, was already the most formidable mathematician of his generation: a child prodigy who published his first paper at eighteen, a founder of game theory, a pioneer of quantum mechanics' mathematical foundations, a central figure of the Manhattan Project who helped design the implosion mechanism of the plutonium bomb. He wrote the EDVAC report after long train conversations with the engineers building ENIAC at the University of Pennsylvania, and though the ideas were partly theirs, his crystalline exposition made the architecture universal; engineers from Manchester to Moscow built 'von Neumann machines' from his description. At Princeton's Institute for Advanced Study he then built the IAS machine, the prototype whose logical design was copied, legally and otherwise, into the first computers of IBM, the RAND Corporation, and half a dozen countries. The architecture had a flaw he acknowledged: the 'von Neumann bottleneck,' the single channel between processor and memory that still throttles computers today. He died of cancer in Washington in 1957, at fifty-three, having also founded the field of cellular automata and sketched the theory of self-reproducing machines. Every smartphone in every pocket is a von Neumann machine, still fetching instructions and data from one memory, still running his 1945 draft. The EDVAC report was only one thread of a mind that seemed to operate in several sciences at once. In 1944 he had published, with the economist Oskar Morgenstern, the 'Theory of Games and Economic Behavior,' founding game theory and giving the Cold War its mathematical language of strategy. On the Manhattan Project he solved the implosion problem, designing the explosive lenses that compressed the plutonium core, work his colleagues considered miraculous. After the war he turned the new computers to science itself: in 1950, on ENIAC, his team ran the first numerical weather forecast, founding computational meteorology. The IAS machine, completed in 1952, was deliberately designed to be copied, and its logical plans were distributed freely; the result was a family of 'IAS machines' at national laboratories and corporations across the world, the direct ancestors of IBM's commercial computers. In his final years, already ill, he wrote about self-reproducing automata, proving mathematically that a machine could contain the instructions to build itself, a result that anticipated both computer viruses and the theory of DNA replication. He served on the Atomic Energy Commission, received the Presidential Medal of Freedom, and died at Walter Reed Hospital in 1957, working, by some accounts, on classified problems to the very end. Princeton's von Neumann, the Hungarians say, was not the smartest man of the century; he was the man the other smart men asked when they were stuck.

Impact on civilization

The stored-program architecture is the reason software exists as an industry. By putting programs into memory alongside data, von Neumann separated the machine from its task: one piece of hardware could become, in succession, a calculator, a typewriter, a telephone exchange, a film studio. That separation created the entire software economy, the app stores, the cloud, and the premise that a computer's value lies not in its circuits but in the infinitely replaceable instructions it holds. It also created computer science as a discipline distinct from electrical engineering, because programs became objects of mathematical study in their own right. The bottleneck he named still shapes chip design eighty years later, which is the mark of an architecture so fundamental that its limitations are laws of nature for engineers. Von Neumann's influence runs through fields that barely touch computers. Game theory reshaped economics, evolutionary biology, and nuclear strategy; his work on the mathematics of quantum mechanics remains the standard formalism taught to physicists; his self-reproducing automata founded the study of artificial life. The stored-program computer is thus only the most visible of his architectures; beneath it lies a habit of mind, find the formal structure, build the universal machine for it, that remade half a dozen sciences.