John Presper Eckert and John William Mauchly developed the first general-purpose electronic digital computer, leading directly to modern methods of computation. While the computer was extremely primitive by later standards, even the ENIAC’s limitations fueled future advances, as they made clear the initial steps necessary in order to realize computers’ potential.
The ENIAC (Electronic Numerical Integrator and Computer) was the first general-purpose electronic digital computer. By demonstrating the feasibility and value of electronic digital computation, it initiated the computer revolution. The ENIAC was developed during World War II at the Moore School of Electrical Engineering
Even early in the war, the BRL’s needs for ballistic firing tables had already far outstripped the combined abilities of the available differential analyzers (Vannevar Bush’s
The first substantial electronic computer, the ENIAC was designed, built, and debugged within two and one-half years. Even given the highly talented team, such a task could be accomplished only by taking as few design risks as possible. Thus, the ENIAC ended up as an electronic version of prior existing computers: Its functional organization was similar to that of Bush’s differential analyzer, while it was programmed via a plugboard (similar to a telephone switchboard) along the lines of earlier electromechanical calculators created by International Business Machines (IBM). Another consequence of the need to emulate prior devices was that the internal representation of numbers in ENIAC was decimal, rather than the now-standard binary, since the familiar electromechanical computers had all used decimal digits.
Although the ENIAC was completed only after the end of the war, its primary use remained military. In fact, the first production run on the system was a two-month calculation needed for the design of the hydrogen bomb. John von Neumann, working as a consultant both to the Los Alamos Scientific Laboratory and to the ENIAC project, arranged for the production run to be performed immediately prior to ENIAC’s formal dedication in 1946.
The ENIAC was an impressive machine: It contained eighteen thousand vacuum tubes, weighed twenty-seven metric tons, and occupied a nine-by-fifteen-meter room. The device’s final cost to the U.S. Army was about $486,000. In return, the Army received a machine that computed up to one thousand times faster than its electromechanical precursors; for example, addition and subtraction required only two hundred microseconds (millionths of a second). The basic cycle time of the computer was a remarkably fast ten microseconds, which meant that it ran at a rate of 100,000 cycles per second (100 kilohertz). Arithmetic would have been even faster, except that the ENIAC’s decimal implementation required multiple pulses per digit. As a result, an addition operation took twenty cycles (or one-five thousandth of a second) to complete. Nevertheless, at its dedication ceremony, the ENIAC was fast enough to calculate a fired shell’s trajectory faster than the shell itself took to reach its target.
The machine was orders of magnitude more complex than any predecessor and employed a risky, new technology in vacuum tubes
Vacuum tubes perform most reliably when they are warm, so, once it was turned on, the machine was supposed to kept running. A historical oddity is that for some months after the ENIAC’s delivery, the BRL continued its normal practice of shutting off all of its equipment at night. As a consequence, much of the following day was wasted getting the ENIAC operational again. The new speed of the computer created a new problem: It required high-speed availability of both programs and data in order to remain efficient. Otherwise, the computer would spend most of its time idling, waiting for data to be fed into it.
Data for the ENIAC were held in twenty accumulators, each of which was capable of holding ten digits and one operator; in modern terms, this represents a total of only about one hundred bytes of storage. Such limited storage posed a severe difficulty for performing complex calculations. In many cases, the intermediate results of such calculations were actually punched onto IBM cards and fed to later program stages via a card reader—a process that hardly took advantage of the underlying machine speed.
The master control unit supported a variety of useful programming functions, including iterative loops, conditional branches, and subroutines. Programming the ENIAC was accomplished by setting switches and physically connecting accumulators, function tables (a kind of manually set read-only memory), and control units. Connections were made via cables running between plugboards. This was a laborious and error-prone process, often requiring a one-day setup time. The team recognized this problem, and in early 1945, Eckert, Mauchly, and Neumann worked on a design for a new machine. Their basic idea was to treat both program and data as the same kind of object, and in particular to store them in the same high-speed memory; in other words, they planned to produce a stored-program computer. Neumann described and explained this design in his “First Draft of a Report on the EDVAC”
Although Eckert had introduced the concept of a stored-program computer even before Neumann joined the ENIAC project, Neumann contributed new design techniques and provided the first general, comprehensive description of the stored-program
As a consequence of the secrecy requirements that initially prevented Eckert from taking credit for his contribution to stored-program computing, a bitter dispute later arose about who “really” invented the concept of stored-program computers. The answer appears to be that the invention was a joint effort. Special claims to intellectual priority by either Eckert or Neumann are undercut by the fact that Alan Mathison Turing’s
After delivery of the ENIAC, Neumann suggested that the computer could be wired in such a way that a set of instructions would be permanently available and specific instructions could be selected from that set by entries in the function tables. Adele Goldstine
The ENIAC’s electronic speed and the stored-program design of the EDVAC clearly had incredible potential even during the vacuum-tube age. This potential, however, posed a serious engineering challenge, one which had to be solved in order for computing to develop further. It was necessary to produce a computer memory store that would be simultaneously large, inexpensive, and fast. Without such fast memories, the electronic control logic at the heart of a computer would spend most of its time idling. Vacuum tubes themselves (used in the control) were not an effective solution to the problem because of their large power requirements and heat generation.
The EDVAC design draft proposed using mercury delay lines, which were used earlier in radar. These delay lines converted an electronic signal into a slower, acoustic signal in a mercury solution; to provide continuous storage, the acoustic signal picked up at the other end was regenerated and sent back into the mercury. Maurice Vincent Wilkes
Neumann, together with Arthur Walter Burks
In the meantime, Eckert and Mauchly, in a dispute with the Moore School of Electrical Engineering over the ENIAC patent, had quit and formed the Electronic Control Company (later the Eckert-Mauchly Computer Corporation). They managed to keep the patent rights. In 1949, they produced the BINAC computer (at a loss) and the UNIVAC I in 1951, with both machines using mercury storage. They sold out to Remington Rand when they had financial problems, and Eckert ultimately became a vice president of Sperry Rand. The patent issue, however, was not over. After Sperry Rand tried to enforce patent rights, Honeywell sued. In 1973, the patent was invalidated, largely for technical legal reasons, but also because Atanasoff had demonstrated previously the viability of some of the innovations the patent attributed to the ENIAC.
The memory problem that the ENIAC introduced was resolved finally with the invention of the magnetic core in the early 1950’s. Core memory was installed on the ENIAC and soon on all new machines. The ENIAC continued in operation until October, 1955, when parts of it were retired to the Smithsonian Institution. Having proved the viability of digital electronics, and having led directly to stored-program computers, its impact can be recognized in every digital computer today.
-
Burks, Arthur W., and Alice R. Burks. “The ENIAC: First General-Purpose Electronic Computer.” Annals of the History of Computing 3 (October, 1981): 310-399. A technical examination of the origin and design of the ENIAC, with particular consideration of the related patent dispute. Incorporates commentary by other principals and responses by the authors. Usefully illustrated. -
Campbell-Kelly, Martin, and Michael R. Williams, eds. The Moore School Lectures. Cambridge, Mass.: MIT Press, 1985. Lecture notes from a special course held at the Moore School in the summer of 1946 that greatly aided the dissemination of the stored-program concept. Lecturers included all the major designers involved in the ENIAC and many others. Papers range from introductory to highly technical. -
Fleck, Glen, ed. A Computer Perspective. Cambridge, Mass.: Harvard University Press, 1973. An entertaining, pictorial survey of the history of computing machinery from 1890 through the 1940’s, based on an IBM-sponsored exhibition. Contains photographs of many pioneer researchers, their machines, and various related documents. -
Goldstine, Herman H. The Computer from Pascal to von Neumann. Princeton, N.J.: Princeton University Press, 1972. Especially interesting for its detailed account of the developments at the Moore School and the Institute for Advanced Study, in which Goldstine was directly involved. -
Hally, Mike. Electronic Brains: Stories from the Dawn of the Computer Age. Washington, D.C.: Joseph Henry Press, 2005. A survey of the creation of the most important early computers, beginning with the ENIAC. Bibliographic references and index. -
McCartney, Scott. ENIAC: The Triumphs and Tragedies of the World’s First Computer. New York: Walker, 1999. Thorough history of the ENIAC, including philosophical precursors and practical analysis of the contribution of each participant in the program. Bibliographic references and index. -
Metropolis, N., J. Howlett, and Gian-Carlo Rota, eds. A History of Computing in the Twentieth Century. New York: Academic Press, 1980. A collection of papers originally presented at a conference on computing history at the Los Alamos Scientific Laboratory in 1976. The authors include many of the principals involved in the early development of computers and computer software. Annotated bibliography. -
Randell, Brian, comp. The Origins of Digital Computers. New York: Springer-Verlag, 1973. This anthology reprints historic papers from early computer development efforts. Includes a forty-page annotated bibliography. -
Stern, Nancy. From ENIAC to UNIVAC: An Appraisal of the Eckert-Mauchly Computers. Bedford, Mass.: Digital Press, 1981. Reviews the history of Eckert and Mauchly’s association and the computers they developed, concentrating on the social aspects. Issues verdicts on the priority and personal disputes in the story which cannot be assessed in isolation. Includes, as an appendix, Neumann’s “First Draft of a Report on the EDVAC.” -
Williams, Michael R. A History of Computing Technology. Englewood Cliffs, N.J.: Prentice Hall, 1985. A comprehensive and balanced presentation of the development of computer technology from the numerical systems of ancient Egypt through the first generation of electronic digital computers. Recommended for all general readers.
First Electronic Stored-Program Computer Is Completed
UNIVAC I Becomes the First Commercial Electronic Computer