Arthur W. Burks, early computer theorist, dies

Burks was a member of the team that designed the Eniac computer, a frequent collaborator of John von Neumann and a pioneer in computing education. (From The New York Times)

3 min read
Arthur W. Burks, a member of the team that designed the Eniac computer, a frequent collaborator of John von Neumann and a pioneer in computing education, died Wednesday at a nursing home in Ann Arbor, Mich. He was 92.

The cause was complications of Alzheimer's disease, said his daughter, Nancy Burks.

Burks, whose Ph.D. was in philosophy, entered computing in its earliest days and viewed the field as a protean science that had the potential to deeply influence many disciplines.

At the University of Michigan, where he taught for four decades, Burks championed the potential of computing long before there were academic departments in computer science. In 1949, he founded a Logic of Computers Group, and in 1957 he started a graduate program in Communication Sciences.

The program was a blend of computing, language study and biology, which addressed challenges including speech recognition, machine learning and complex adaptive systems, like those later called neural networks.

"Art Burks always viewed computing from a very broad, cross-disciplinary perspective," said John H. Holland, a professor of computer science and psychology at the University of Michigan who was a student of Burks'.

Burks was an important contributor to the theoretical side of computer science rather than the engineering side, said William Aspray, a historian at Indiana University. Burks' legacy, Aspray said, can be seen in the current interdisciplinary research into complex adaptive systems at the Santa Fe Institute, the Massachusetts Institute of Technology and the Center for the Study of Complex Systems at the University of Michigan.

Burks was born on Oct. 13, 1915, in Duluth, Minn., and his family moved to Chicago when he was 10. His father, Walter, was a high school mathematics teacher, and his mother, Cora, had a college degree. Burks was one of four children, all now deceased.

Burks earned his doctorate in philosophy at the University of Michigan in 1941, and his particular focus was symbolic logic, which is highly mathematical. After he received his doctorate, Burks saw a newspaper advertisement that said the Moore School of Engineering at the University of Pennsylvania was offering crash courses in electronics for qualified candidates, who would then work on an unnamed wartime project for the government, recalled his wife, Alice R. Burks.

Burks was accepted and began working on the Eniac, short for Electronic Numerical Integrator and Computer. The Eniac's job was to calculate the trajectory of artillery shells. It was finally finished and demonstrated in February 1946, too late for use in World War II, but it is credited with being the first electronic digital computer.

It was on the Eniac project that Burks met and began working with von Neumann, the eminent mathematician and computing theorist. von Neumann, Burks and Herman H. Goldstine, who later became a researcher at IBM, worked mostly on the logic and design issues. But two other scientists at the University of Pennsylvania, John W. Mauchly and J. Presper Eckert, were the engineering wizards who built the Eniac.

The more lasting contribution to computing of the theorists--von Neumann, Burks and Goldstine ? came in two papers published in 1945 and 1946. The first was a draft of the design for the Eniac's successor, the Edvac, and the second paper was presented at the Institute for Advanced Study at Princeton. The first was written by von Neumann, and the second listed all three men as co-authors. The papers described the design for the modern programmable, general-purpose computer, initially called the "stored-program computer," and later termed the von Neumann architecture.

Burks continued to collaborate with von Neumann, and his most significant contribution came after von Neumann died in 1957. Burks completed and edited a paper von Neumann had been working on, "Theory of Self-Reproducing Automata," which described how a machine might reproduce itself; it was published in 1966.

"That paper had a huge impact, not only in computing but in biology and philosophy as well," said Holland of the University of Michigan. "Until then, it was assumed that only living things could reproduce."

"The key ideas were all von Neumann's," he added. "But there were few people in the world besides Art Burks who could have completed that paper."

Burks is survived by his wife of 65 years. The couple met in July 1942 at the Moore School, where she was doing calculations for the Eniac, she said. They were married less than a year later.

Burks is also survived by his three children, Nancy, Edward and Douglas, all of Ann Arbor.