A Brief History of Digital Computers

Bruce Gillespie, March 2002
B Sc Elec Eng (Wits), M Sc History of Science, Technology and Medicine (Imperial)

Although both electrical and mechanical calculating and computing machines were in wide use since the beginning of the 20th century, digital computing was a unique departure. The innovation of information being represented using only two states (On or Off) came to be known as "Digital", whereby a Base 2 numbering system was used. Binary Arithmetic and Logic provided the tools for these machines to perform useful functions. By using only two states, engineering was greatly simplified, and universality and accuracy increased. Further developments from the early purpose built machines, to ones that were programmable accompanied by many key technological developments, resulted in the well-known success and proliferation of the Digital Computer.

Amongst the first recorded digital computers was a machine built by a German, Konrad Zuse, In 1936 he constructed a binary calculating machine out of electromechanical relays in his living room. This performed binary arithmetic, floating point calculations and program control was by a punched tape, widely used in telegraphs. His work was independent of projects elsewhere in the UK and USA, but did not result in proliferation nor market success.

The first round of digital computers in the USA also made use of this electromechanical relay technology, and was seeded by Claude Shannon's 1937 MIT Master thesis. He described a way of using symbolic logic to improve electrical switching circuits and described an electrical circuit made out of relays and switches that could add binary numbers. Since relay technology was widely used in telephone exchanges, it was mature and broadly available. Relays operate on pulses of fixed voltage electricity and are either on or off, thus facilitating the two binary states. Bell Telephone Laboratories in the USA tested ideas on a small scale with their Relay Interpolator, which consisted of 500 relays and later the Ballistic Computer which contained 1300 relays. Their work cumulated in 1944 with an all-purpose computer with 9000 relays and operated by 50 teletype consoles.

International Business Machines (IBM), which had established itself in the office adding and tabulating machine market, took up ideas put forward by Harvard University's Howard Aiken to solve existing problems of "insufficient means of mechanical computation". Their ‘Mark 1' computer was completed in 1944 and also used electromechanical relay technology. It was considered the first automatic general purpose digital machine. It had four tape readers in which data and program code was fed, and subroutines stored. A more capacious Mark II followed. The machines were used by the US Navy for tackling ballistic problems, one of the recurring needs during wartime for ‘number crunching'.

In the UK pioneering efforts in digital computing were made by mathematician Alan Turing, who's idea of a "Universal Turing Machine" helped break the conceptual ground for automated computers tackling mathematical problems. His 1936 paper "On Computable Numbers" theoretically described a machine that could do any calculation that could be done by a human. During the Second World War, his work with developing computing systems at Bletchley Park was fundamental in cracking German U-boat "Enigma"communications and keeping the Atlantic shipping lanes open, ultimately contributing towards the Allied victory. The purpose-built 1944 ‘Colossus' was built out of relays from British Post Office telephone exchanges, and operated with a fixed sequence for deciphering the Germans U-Boat communications encrypted with Enigma codes.

The development and application of the electronic vacuum tube (also known as valves), as a switching unit in digital computers permitted vastly improved performance compared to relay technology. The first unit to employ large scale use of valves for digital computing was the ENIAC (Electronic Numerical Integrator And Computor) which consisted of 18 000 valves. Built by the Moore School of Electrical Engineering at the University of Pennsylvania to compute ballistic tables for the US Army, it was only completed after the close of the war in late 1945. Although it could only store 20 binary words, it's unprecedented calculation speed attracted attention from scientists and professionals, in particular a key member of the Manhattan Atomic Bomb project, mathematician John von Neuman. It was used for other scientific purposes such as weather predicting and other fluid dynamic calculations, including calculations for the fusion Hydrogen Bomb. However programming the ENIAC meant setting arrays of ‘function' switches and a maze of patch cables. This was time consuming and meant that it could not be used for computing whilst it was being programmed. This drove the development of the stored program computer, whereby the computational instructions could be prepared on tape and loaded quickly into memory.

The EDVAC (Electronic Discrete Variable Calculator) was the first stored program electronic digital computer and was also built by the Moore School. It was highly influenced by von Neumann work in formal logics since he provided critical direction into the logical structure and design of the computer. He abstracted the logical design from the hardware, and introduced the concept of the ‘stored program' whereby programming instructions are stored as numbers in memory alongside numerical data. The EDVAC also introduced mercury delay lines for primary memory. Developed by John Eckert for the project, these devices increased storage density by a factor of one hundred over valve devices.

The ideas and work of the Moore School rapidly proliferated there after. A course on computer design based on the von Neuman concept was run by that institution in 1946, and was attended by British and American professionals. Also in 1946, von Neuman, Herman Goldstine and Arthur Burks published a comprehensive report of their work at Princeton's Institute of Advanced Study Electronic Computer Project where they had established themselves following their involvement with the Moore School. The report detailed the operation and architecture of their work on digital computers, and has been described as the blue print for ‘modern' digital computing. The authors insistence that it be placed in the public domain, as is the tradition with scientific publishing, was certainly an important factor in the spread and adoption of this knowledge.

The Whirlwind was developed at the Servomechanism Laboratory at MIT in Boston, and was sponsored by the US military for real-time control and flight simulation work. The project began in 1946 and whilst it drew from the work at Princeton and Moore Schools work, the designers background in automatic control and electrical engineering gave it a certain technological uniqueness. The Whirlwind project resulted in some key technological developments such as magnetic core memory and the parallel synchronous method of handling information. Developed by Jay Forrester in 1949, magnetic core memory was an important step in miniaturisation and speed of computer memory. The Whirlwind was only commissioned in 1958, finding use as the control and prediction computer for the SAGE air defence system. This relied on real-time communications to remote radar stations for which a key digital communication technology was developed: the modem. However it has been argued that Whirlwind's time was over before it ever began: the system had become redundant by the Cold War proliferation of intercontinental ballistic missiles.

The UNIVAC is regarded as the first commercial digital computer. It was designed and built by J. Presper Eckert and John Mauchly who had worked extensively on the engineering of ENIAC project. The acronym stems from "Universal Automatic Computer", the intention being that it would have universal appeal to scientists, engineers and business. Although the core functional ‘organs' were similar to the EDVAC, it was a more robust machine and required less maintenance. The UNIVAC incorporated some key improvements which vastly speeded up data processing which included a magnetic tape system for secondary memory storage and a data buffering mechanism to the delay line primary memory storage. The first commercial installation of a UNIVAC was at the U.S. Census Bureau in 1951.

IBM responded in 1952 with the Model 701, which was technologically similar to the UNIVAC with the exception that it featured cathode ray tube memory. This was a technology appropriated from the television industry, the phosphorescence of the screen providing the ability to store binary signals. Other key feature was the use of a magnetic oxide coated drum for secondary storage (the predecessor of the ‘hard disk') as well as a light-weight plastic tape tertiary memory. This had a lower inertia than the metal tapes of the UNIVAC and sped up tape operations considerably. Due to IBM's massive market presence and infrastructure, the 701 was a considerable market success.

Following the success and proliferation of these computers in business and scientific research organisations from the early 1950's, a plethora of startups presented the global market with a diverse range of digital computers. These companies were initiated by key staff members from the pioneering organisations, such as Seymour Cray and William Norris. Besides technological improvements in processing speed, primary and mass storage memory density, architecturally digital computers did not diverge from the classic ‘von Neuman' architecture. Further military-sponsored research in the USA produced key advances such as the transistor, the integrated circuit and computer networking (from which The Internet grew) although Personal Computing was more of a civilian initiative. Larger demand for computers in all spheres of society spawned a global industry which ultimately drove down prices and increased miniaturisation, and resulted in the digital computer becoming ubiquitous in most industrialised societies by the turn of the century.

Bibliography and futher reading:

William Aspray, "John von Neuman and the Origins of Modern Computing", (MIT Press, 1990)

Paul E. Ceruzzi, "A history of modern computing", (MIT Press, 1998)

Charles and Ray Eames, "A computer perspective, background to the computer age", (Harvard University Press, 1973)

Herman H. Goldstine, "The Computer, from Pascal to von Neuman", (Princeton University Press, 1972)

Andrew Hodges, "Alan Turing, the enigma", (Burnett Books, 1983)

Thomas P. Hughes, "Rescuing Prometheus, Four monumental projects that changed the modern world", (Vintage Books, 1998)