Chapter 2 - Background
2.15 The Emergence of First Generation Computers 1946-1959
In late 1945 researchers at the Moore School of the University of Pennsylvania powered up a machine that was 100 feet long, 10 feet high, and 3 feet deep. It contained 17,000 vacuum tubes, about 70,000 resistors, 10,000 capacitors, and 6,000 switches. The Electronic Numerical Integrator and Calculator (ENIAC) was the first fully electronic computer. Designed and developed by J. Presper Eckert and John W. Mauchly under contractor supervision of Lieutenant Herman Goldstine,25 ENIAC funding came from the Army’s Ballistic Research Laboratory at Aberdeen. ENIAC had a serious defect. Although it could compute several hundred times faster than an electro-mechanical or relay-type machine26 , the computer required rewiring with each new problem, consuming from 30 minutes to a full day.27 Nevertheless, in instantiating the first electronic computer, Eckert, Mauchly and Goldstine advanced the trajectory of computers beyond being just an idea.
Upon completing the design of ENIAC, the team faced a new challenge: to design the next computer, one to be significantly better, preferably without the impossible constraint of rewiring. In the summer of 1944, Lieutenant Goldstine, by chance, encountered John von Neumann while waiting for a train. Goldstine began to tell von Neumann, the foremost applied mathematician of his time, about ENIAC. Von Neumann was himself involved in a number of secret projects needing more computation than then available, yet did not know of ENIAC.
On August 7, 1944, von Neumann visited the Moore School and began contributing ideas immediately. The collaboration between von Neumann, Eckert, Mauchly and Goldstine proved fruitful when they hit upon the concept of storing the logic instructions in memory – the “stored-program”28 computer. Instead of manual resetting of the switches, or worse yet, rewiring, to set-up the calculations of a new problem, the programmer could modify the program arithmetically.29 With these new architectural ideas, they designed the ENIAC’s successor, the EDVAC (Electronic Discrete Variable Automatic Calculator).
Elsewhere, Thomas J. Watson Jr. had rejoined IBM after serving in the Air Force. In early 1946, Eckert and Mauchly gave Watson Jr. a tour of the Moore School and ENIAC. Although Watson Jr. sensed that Eckert and Mauchly thought they had a product with which to best IBM, he was unimpressed. “The truth was that I reacted to ENIAC the way some people probably reacted to the Wright brother’s airplane: it didn’t move me at all. I can’t imagine why I didn’t think, “Good God, that’s the future of the IBM company.”30
Funding the development of computer systems, such as the ENIAC and EDVAC, represented only one way the government advanced understanding of computer technologies. Government agencies also sponsored meetings and courses, with the objectives of theory development, diffusion of knowledge, and training of personnel needed to explore the computer trajectory. Diffusing the growing knowledge and practice of computers ensured the government a future base of personnel to design and build the computers the military would need.
During the summer of 1946, the Moore School held a six-week course entitled, “Theory and Techniques for the Design of Electronic Digital Computers,” sponsored by the Office of Naval Research and the Army Ordinance Department.31 Six months later, the Navy sponsored a four-day conference at Harvard, 350 people attended.32 Representatives of government agencies, universities and a wide range of companies attended both of these significant events. Although many companies were learning about computers, none of them would be the first to act; that honor belongs to a start-up formed by Eckert and Mauchly in March 1946.
In 1946, the University of Pennsylvania dismissed Eckert and Mauchly because of their interests in commercializing the ENIAC and EDVAC. The two perceived an economic opportunity in selling computers and together formed the Electronic Control Company. The Census Bureau awarded them their first contract in June, beating out Raytheon. In 1947, the company changed its name to the Eckert-Mauchly Computer Corporation.
Two additional contracts to build computers were signed – with A.C. Nielson and Prudential Life Insurance Company – with badly needed cash advances to fund continuing product investment. Henry Straus, a Delaware racetrack owner and vice president of American Totalizer, committed to invest half a million dollars for 40 percent of Eckert-Mauchly common stock. But Straus died in an airplane crash, and local financial institutions refused to honor the notes Straus had given the company. By 1949, the company verged on bankruptcy, despite having their three contracts.33
Eckert and Mauchly contacted everyone they knew who might be interested in either funding them, or acquiring them: NCR, Remington Rand, IBM, Philco, Burroughs, Hughes Aircraft, and others. They signed their acquisition by Remington Rand in February 1950. Remington Rand then attempted to cancel the three contracts. Nielson and Prudential agreed reluctantly. But the Census Bureau refused to cancel. They were going to make Remington Rand deliver the agreed upon computer. This act, obviously, influenced the economic history of computers by forcing product instantiation.
In the Spring of 1951, the Eckert-Mauchly division of Remington Rand shipped the first UNIVAC I to the Census Bureau. The next five UNIVAC I’s shipped to the government as well, to the Atomic Energy Commission (2), Air Force, Army, and Navy.34
Another start-up, however, captured the honors of shipping the first commercial computer. The founders of Engineering Research Associates (ERA), after resigning from the Naval Communications Supplementary Activity, signed a contract with the Navy to develop the Atlas I computer. They had the right to resell the same technology commercially. The Atlas I computer as the ERA 1101 shipped in December 1950.
In 1952, Remington Rand, repeating a pattern of organizational development used in the office machinery, acquired their second computer firm, ERA. This time Remington Rand owned the two leading firms, not firms wanting to be acquired because they had trouble competing. Remington Rand controlled the emerging computer market-structure, yet remained a distant second to IBM in office machinery, especially tabulating equipment. In 1954, Remington Rand sold their first UNIVAC to a commercial customer: General Electric.
- [25]:
Goldstine held a Ph.D. and participated actively in the project.
- [26]:
RichardR. Nelson, “The Computer Industry,” Government and Technical Progress: A Cross-Industry Analysis,” Pergammon Press, 1982, p. 165
- [27]:
George Schussel, “IBM and REMRAND” Datamation, May 1965, p. 55
- [28]:
Stored-programing laid the beginnings of software, both as component and architecture.
- [29]:
Nelson, p. 167
- [30]:
Thomas J. Watson Jr., “Father, Son & Co.,” Bantam Books, 1990, p.143
- [31]:
Nelson, p 167
- [32]:
Ibid.., p 168
- [33]:
Nelson, pp 169-170
- [34]:
Nelson, p. 170