The BINAC demonstrated the viability of several engineering innovations—most important, the stored-program concept—that became standard features of later computers.
The evolution of computers lies at the intersection of the inventive enthusiasm of mathematicians and engineers and the practical computational needs of various segments of society. The former has been responsible for designs; the latter have motivated the funds necessary to realize those designs. In the 1820’s, the need for error-free mathematical and astronomical tables for use in navigation, unreliable versions of which were being produced by human “computers,” stirred Charles Babbage to design and partially construct some of the earliest prototypes of modern computers, with substantial but still inadequate funding from the British government. In the 1880’s, the search by the United States Bureau of the Census for a more efficient method of compiling the 1890 census led Herman Hollerith
John William Mauchly at the central control of BINAC, 1948.
The emergence of modern electronic computers similarly began in response to a practical problem. During World War II, there was an urgent need in the American military for reliable and quickly produced mathematical tables to be used in aiming various types of artillery. The calculation of very complex tables had progressed somewhat since Babbage’s day, and the human computers were being assisted by mechanical calculators. The growing demand for increased accuracy and efficiency was pushing the limits of these machines, and in 1946, following three years of intense work at the University of Pennsylvania’s Moore School of Engineering, John Presper Eckert and John William Mauchly presented their solution to the problem in the form of the Electronic Numerical Integrator and Computer (ENIAC)
The ENIAC, built under a contract with the Army’s Ballistics Research Laboratory, was a great success for Eckert and Mauchly, but even before it was completed, they were setting their sights on loftier targets. The primary drawback of the ENIAC was the great difficulty involved in programming it. Whenever the operators needed to instruct the machine to shift from one type of calculation to another, they had to reset a vast array of dials and switches, unplug and replug numerous cables, and make various other adjustments to the multiple pieces of hardware involved. Such a mode of operation was deemed acceptable for the ENIAC because, in computing firing tables, it would need reprogramming only periodically. If the instructions could be stored in a machine’s memory along with the data, however, such a machine would be able to handle a wider range of calculations with ease and efficiency.
The idea of a stored-program computer first appeared in a paper written by the English mathematician Alan Mathison Turing and published in London in 1937. In this paper, Turing described a hypothetical machine of quite simple design that could be used to solve a wide range of logical and mathematical problems. A most significant aspect of this imaginary Turing machine is that the tape that would run through it would contain both information to be processed and instructions on how to process it. The tape would thus be a type of memory device, storing both the data and the program as sets of symbols that would be dealt with indifferently by the machine. Turing never attempted to construct this machine, and it was not until 1946 that he developed a design for an electronic stored-program computer, a prototype of which was built in 1950.
In the meantime, John von Neumann, a Hungarian mathematician acquainted with Turing’s ideas, joined Eckert and Mauchly in 1944 and contributed to the design of ENIAC’s successor, the Electronic Discrete Variable Automatic Computer (EDVAC)
In March of 1946, Eckert and Mauchly, frustrated by a controversy over patent rights for the ENIAC, resigned from the Moore School and several months later formed the Philadelphia-based Electronic Control Company, with a contract from the National Bureau of Standards and the Census Bureau to build a much grander computer called the Universal Automatic Computer (UNIVAC)
Building the UNIVAC, however, proved to be much more involved and expensive than anticipated, and the funds provided by the original contract were inadequate. Eckert and Mauchly, therefore, took on several other smaller projects in an effort to raise funds. On October 9, 1947, they signed a contract with Northrop Aircraft Company of Hawthorne, California, to produce a relatively small computer to be used in the guidance system of a top-secret missile called the Snark, which Northrop was building for the Air Force. This computer, known as the Binary Automatic Computer (BINAC), turned out to be Eckert and Mauchly’s first commercial
The BINAC was designed to be at least a preliminary version of a compact, airborne computer and was somewhat diminutive for its time. It had two main processing units. These contained a total of fourteen hundred vacuum tubes, a drastic reduction from the eighteen thousand used in the ENIAC. There were also two memory units, as well as two power supplies, an input converter unit, and an input console, which used either a typewriter keyboard or encoded magnetic tape (the first time such tape was used for computer input). Because of its dual processing, memory, and power units, the BINAC was actually two computers, each of which would continually check its results against those of the other in an effort to identify errors. If a discrepancy was found, the computer would automatically stop. If the results matched, it would continue its calculations.
The memory units contained an important innovation: the mercury delay line. This device, conceived by Eckert several years earlier and developed for use in the EDVAC, consisted of a mercury-filled tube through which electrical impulses (transformed into sound waves) passed much more slowly than through a wire. This delay made it possible for impulses, carefully quantified to represent the binary digits 1 or 0, to be stored or maintained electronically in the tube almost indefinitely. The mercury delay line greatly increased the machine’s memory efficiency and capacity and made it possible to store not only data but also instructions within the computer. It was, therefore, closely tied to the successful implementation of the stored-program concept.
Despite its much smaller size, the BINAC had an internal memory capacity many times greater than the ENIAC, whose performance it could at worst approximate and at best far surpass. The BINAC was capable of executing thirty-five hundred additions or subtractions per second (compared to ENIAC’s five thousand) and one thousand multiplications or divisions per second (exceeding ENIAC’s 333 multiplications and 41 divisions).
The BINAC was originally scheduled for completion on May 15, 1948, seven months after the contract was signed. Because of a variety of administrative, financial, and engineering problems, however, it did not become operational until August of 1949. Because Eckert and Mauchly wanted to publicize the fact that their company had now produced a workable computer, public demonstrations of the BINAC were held in Philadelphia from August 18 through August 20. They took this opportunity to advertise their much larger and concurrent project, the UNIVAC, to the press and potential customers. The National Bureau of Standards and the Census Bureau had sent a delegation to these demonstrations, and the successful operation of the BINAC convinced these agencies that sufficient progress was being made, enabling Eckert and Mauchly to receive additional payments on the UNIVAC contract.
On August 22, Eckert and Mauchly held the BINAC demonstration test for representatives of Northrop Aircraft Company. It performed admirably for a period of seven hours and ten minutes, with forty minutes of downtime for repairing or replacing parts. Northrop was satisfied with the performance, although not with the late date on which it occurred, and in September shipped the computer to their headquarters in California.
The contract for the BINAC had specified a total cost of $100,000, with $80,000 being paid upon signing the contract and the remaining $20,000 due upon finishing the job. The actual cost, however, turned out to be $278,000, a discrepancy indicative of Eckert and Mauchly’s inability to provide accurate cost estimates. Despite the fact that the completion of the BINAC brought much-needed capital into their company, as well as the fact that they had several UNIVAC contracts, their fledgling enterprise with more than one hundred employees was sinking into debt. In February of 1950, Eckert and Mauchly agreed to sell their corporation to Remington Rand, representatives of which had attended the BINAC demonstrations. Remington Rand became Sperry Rand a few years later and produced UNIVACs, under the direction of Eckert and Mauchly, with great commercial success.
Throughout its construction, the BINAC stood in the shadow of the more powerful UNIVAC. Mauchly had suggested in 1948 that the company should concentrate on producing several BINACs to sell, but Eckert and the engineering staff persuaded him to follow their preference for the UNIVAC. The importance of the BINAC for the company was seen, at least in part, in terms of how it contributed to the development of the UNIVAC. It did, indeed, contribute in significant ways.
The fact that the BINAC was relegated to a secondary status meant that it did not receive the attention of the best engineers and also that some of the parts used in its construction were not of the highest quality. These factors, as well as the fact that the machine emerged from the trip to California in a battered condition, resulted in a poor performance record at Northrop.
It is the design embodied in the BINAC that is the real source of its significance. It demonstrated successfully the benefits of a dual-processor design for minimizing errors, a feature adopted in many subsequent computers. It showed the suitability of magnetic tape as an input-output medium. Its most important new aspect was its ability to store programs in its relatively spacious memory, the principle that Eckert, Mauchly, and von Neumann had originally designed into the EDVAC. In this respect, the BINAC was a direct descendant of the EDVAC.
The stored-program feature of BINAC’s design greatly enhanced its efficiency. It contributed to its speed, because the electronically stored instructions could be accessed as quickly as the electronically encoded information. The stored-program feature also accounted for the ease with which the BINAC could be reprogrammed, and thus for its flexibility in solving a variety of computational problems. In short, the stored-program principle gave electronic computers new powers, quickness, and automatic control that, as they have continued to grow, have contributed immensely to the aura of intelligence often associated with their operation.
The BINAC successfully demonstrated some of these impressive new powers in August of 1949 to eager observers from a number of major American corporations. It helped convince many influential leaders of the commercial segment of society both of the engineering expertise of Eckert and Mauchly and of the promise of electronic computers in general. In doing so, it helped ensure the further evolution of computers.
-
Augarten, Stan. Bit by Bit: An Illustrated History of Computers. New York: Ticknor & Fields, 1984. A well-written and pleasingly presented treatment of the development of modern computers and their predecessors from antiquity to the twentieth century. Chapter 5 is devoted to stored-program computers such as the BINAC, EDVAC, EDSAC, and UNIVAC, and describes the various contributions of Turing, von Neumann, Eckert, and Mauchly. Thoroughly illustrated, with a bibliography and index. -
Davis, Martin. The Universal Computer: The Road from Leibniz to Turing. New York: W. W. Norton, 2000. Rather than focus on the physical machines and engineering feats of their inventors, this text looks at the philosophical history of the idea of the computer and of the mathematics behind computing—and the relationship between those theoretical developments and the actual machines built in the 1940’s and 1950’s. Bibliographic references and index. -
Goldstein, Herman H. The Computer from Pascal to Von Neumann. Princeton, N.J.: Princeton University Press, 1972. Goldstein worked with Eckert, Mauchly, and von Neumann on the ENIAC and EDVAC, and he discusses these machines in depth. Places particular emphasis on the contributions to electronic computing made by von Neumann. Weaves together technical details and personal reminiscences. -
Norberg, Arthur L. Computers and Commerce: A Study of Technology and Management at Eckert-Mauchly Computer Company, Engineering Research Associates, and Remington Rand, 1946-1957. Cambridge, Mass.: MIT Press, 2005. Delves into the inventions of Eckert and Mauchly once they left the Moore School, as well as the commercial considerations that shaped the development of both their technological innovations and their corporate experience. Bibliographic references and index. -
Shurkin, Joel N. Engines of the Mind. New York: W. W. Norton, 1984. Chronicles the development of computers from the ancient world to the modern age. About two-thirds of the book discusses the work of Eckert and Mauchly in the 1940’s and 1950’s; devotes attention to the controversy surrounding the claim that John V. Atanasoff invented the first electronic computer at Iowa State University in 1939. Includes a glossary, bibliography, and detailed index. -
Stern, Nancy. From ENIAC to UNIVAC: An Appraisal of the Eckert-Mauchly Computers. Bedford, Mass.: Digital Press, 1981. A thorough and balanced study of the technical, financial, administrative, and intellectual aspects of Eckert and Mauchly’s endeavors. Includes as an appendix the famous “First Draft of a Report on the EDVAC,” written by von Neumann in 1945, which was responsible for von Neumann getting primary credit for designing the first stored-program computer. -
Wulforst, Harry. Breakthrough to the Computer Age. New York: Charles Scribner’s Sons, 1982. A popularly written, anecdotal account of the first electronic computers, primarily the ENIAC, BINAC, and UNIVAC, by a former director of public information for Sperry Univac. Emphasizes the economics, politics, and personalities involved in these projects, rather than the technical aspects. Illustrated, with a brief bibliography and index.
Eckert and Mauchly Develop the ENIAC
UNIVAC I Becomes the First Commercial Electronic Computer