George Stibitz

George Robert Stibitz (April 30, 1904 – January 31, 1995) was an American researcher at Bell Labs who is internationally recognized as one of the fathers of the modern digital computer.

He received a bachelor's degree in mathematics from Denison University in Granville, Ohio, a master's degree in physics from Union College in 1927, and a Ph.D. in mathematical physics from Cornell University in 1930 with a thesis entitled "Vibrations of a Non-Planar Membrane."

He was known for his work in the 1930s and 1940s on the realization of Boolean logic digital circuits using electromechanical relays as the switching element.

Stibitz was born in York, Pennsylvania, the son of Mildred Murphy, a math teacher, and George Stibitz, a German Reformed minister and theology professor.

Throughout his childhood, Stibitz enjoyed assembling devices and systems, working with material as diverse as a toy Meccano set or the electrical wiring of the family home.

In November 1937 he completed a relay-based adder he later dubbed the "Model K" (after his kitchen table, on which it was purportedly assembled), which calculated using binary addition.

Replicas of the "Model K" now reside in the Computer History Museum, the Smithsonian Institution, the William Howard Doane Library at Denison University and the American Computer & Robotics Museum in Bozeman, Montana.

Bell Labs subsequently authorized a full research program in late 1938 with Stibitz at the helm.

He led the development of the Complex Number Calculator (CNC), completed in November 1939 and put into operation in 1940.

Employing electromagnetic relay binary circuits for its operations, rather than counting wheels or gears, the machine executed calculations on complex numbers.

In a demonstration at the meetings of the American Mathematical Society and Mathematical Association of America at Dartmouth College in September 1940, Stibitz used a modified teletype to send commands over telegraph lines to the CNC in New York.

This was the first real-time, remote use of a computing machine.

Stibitz began working at Bell Labs after his doctorate, where he would remain until 1941.

After the United States entered World War II in December 1941, Bell Labs became active in developing fire-control devices for the U.S. military.

The Labs' most famous invention was the M-9 Gun Director, an ingenious analog device that directed anti-aircraft fire with uncanny accuracy.

Stibitz moved to the National Defense Research Committee, an advisory body for the government, but he kept close ties with Bell Labs.

For the next several years (1941–1945), with his guidance, the Labs developed relay computers of ever-increasing sophistication.

The first of them was used to test the M-9 Gun Director.

Later models had more sophisticated capabilities.

They had specialized names, but later on, Bell Labs renamed them "Model II", "Model III", etc., and the Complex Computer was renamed the "Model I".

All used telephone relays for logic, and paper tape for sequencing and control.

In April 1942, Stibitz attended a meeting of a division of the Office of Scientific Research and Development (OSRD), charged with evaluating various proposals for fire-control devices to be used against Axis forces during World War II.

Stibitz noted that the proposals fell into two broad categories: "analog" and "pulse".

In a memo written after the meeting, he suggested that the term "digital" be used in place of "pulse", as he felt the latter term was insufficiently descriptive of the nature of the processes involved.

In the very same moment, he also pointed to the limits of this opposition between analog and digital.

He presented it as a rather theoretical opposition with no practical use, as most computers of the time would consist of both analog and digital mechanisms.

The "Model V", was completed in 1946 and was a fully programmable, general-purpose computer, although its relay technology made it slower than the all-electronic computers then under development.

The application of Math via a machine was an extraordinary effort of Stibitz with Claude Shannon, and is easily understated.

Decimal numbers were encoded as groups of two and five relays, somewhat like the beads on a Chinese abacus.

This biquinary code allowed for elaborate error checking, which ensured that he machine would stop and alert an operator before ever delivering a wrong answer.

Relay computers, unlike their electronic counterparts, had to have error-detecting circuits because a relay can fail intermittently, usually when a piece of dust interferes with a few contact cycles before being dislodged.

Such intermittent errors would have been almost impossible to detect without some sort of internal redundancy.

By contrast, vacuum tubes failed catastrophically, with a resulting computer failure obvious to its operators.

See https://dl.acm.org/doi/pdf/10.5555/1074100.1074162 Even today AI Artificial intelligence makes many mistakes, called "uncertain reasoning" and error detection is much needed.

The Bell Labs computers were powerful, reliable, and balanced machines.

They often outperformed their vacuum tube contemporaries in solving problems for which slower speed was not decisive.

But once the von Neumann-inspired notions of computer architecture became known and accepted, that advantage was lost, as designers elsewhere learned to build electronic computers with none of the architectural drawbacks suffered by machines like the ENIAC.

Thus, the Bell Labs machines represent an evolutionary dead end, although their contribution to the mainstream history of digital computing was profound.

At the end of the war, Stibitz did not return to Bell Labs, but went into private consulting work.

From 1964 until his retirement in 1974, Stibitz was a research associate in physiology at the medical school of Dartmouth College.