George Stibitz and the Complex Number Calculator

Abstract

George Robert Stibitz (1904–1995), a mathematician at Bell Telephone Laboratories, built a binary adder from telephone relays on his kitchen table in 1937. Bell Labs let him turn it into the Complex Number Calculator, which went into service in January 1940, and in September 1940 he operated it from a teleprinter at Dartmouth College in New Hampshire over a telegraph line to New York, the first time anyone used a computer remotely. During the war Bell Labs built five more relay calculators under his guidance; the last of them were general-purpose, programmable machines.

The Kitchen Table

Stibitz was born on 30 April 1904 in York, Pennsylvania, the son of a theology professor and a mathematics teacher, and grew up in Dayton, Ohio, where he attended a technical high school founded by the General Motors engineer Charles Kettering, inventor of the electric starter. He took a bachelor’s degree in mathematics at Denison University, a master’s in physics at Union College in 1927, and a doctorate in mathematical physics at Cornell in 1930, and then joined Bell Telephone Laboratories, the research arm of AT&T, in New York (see Bell Labs: The Idea Factory).

In November 1937 he wired two telephone relays and two light bulbs on his kitchen table into a circuit that added two binary digits and showed the result in the lamps, later known as the “Model K.” Late in 1938 Bell’s management, whose engineers spent much of their time calculating with complex numbers, authorised him to build a full machine.

The Complex Number Calculator

The Complex Number Calculator was assembled by October 1939 and in operation on 8 January 1940. Its more than 400 relays could add, subtract, multiply and divide two complex numbers, numbers of the form a + ib that electrical engineers use for alternating currents. It worked in decimal, but stored each digit N as the binary form of N + 3, a scheme now known as excess-3 or the Stibitz code. Operators worked it from modified teleprinters; the arithmetic unit sat in Bell’s building in New York.

Dartmouth, September 1940

From 9 to 12 September 1940, a mathematical summer meeting brought several hundred mathematicians to Dartmouth College in Hanover, New Hampshire, John von Neumann and Norbert Wiener among them. Bell Labs set up one of the calculator’s teleprinters in a college building and connected it by telegraph line to the machine in New York, about 350 kilometres away. Stibitz described the machine in a talk, and problems typed in at Hanover came back as printed results. It was the first time a computer had been operated from a distance, a quarter-century before time-sharing made remote terminals routine (see Data Communication Prehistory).

The machine was roughly contemporary with Konrad Zuse’s relay-based Z2 of 1939 (see Konrad Zuse and the Z3). The two men were later honoured together: in 1965 they shared the Harry H. Goode Memorial Award.

The Bell Relay Machines

In 1941 Stibitz left Bell Labs for the National Defense Research Committee, which paid for the next four Bell machines; the Complex Number Calculator was retroactively named Model I. Model II was the “Relay Interpolator”; Models III and IV were ballistics computers for the Army and Navy. Model V, completed in 1946, was a fully programmable, general-purpose computer; two copies were built, one for the Army’s Aberdeen Proving Ground and one for the aeronautics laboratory at Langley Field, where it reduced wind-tunnel data. The two Model Vs weighed about ten tonnes each and held 9,000 relays. In the account of the Bell engineer E. G. Andrews, the slow Model V at Aberdeen did more work than ENIAC. A sixth model was finished in 1950 without Stibitz’s involvement; it later went to Brooklyn Polytechnic Institute and then to the Bihar Institute of Technology in India.

The Bell machines were among those presented at the Harvard symposium of January 1947, where Stibitz spoke on the day devoted to logic (see Howard Aiken and the Harvard Mark I); he wrote his own history of them in 1967.

Later Years

After the war Stibitz worked as an independent consultant. In 1964 he joined Dartmouth Medical School to apply computing to physiology; the HNF dates his teaching there 1964 to 1983, while Wikipedia gives his post as research associate as 1964 to 1974. He received the IEEE Computer Pioneer Award in 1982 and entered the National Inventors Hall of Fame in 1985. He died in Hanover on 31 January 1995, and a plaque at Dartmouth marks the 1940 demonstration.