ENIAC or ABC?

The review by J. Ross Macdonald and Harvey G. Cragon (PPhysics Today, July 2000, page 58 ) of ENIAC: The Triumphs and Tragedies of the World’s First Computer seems to be an able assessment of the book and its content. However, perhaps due to misstatements in the book, the review fails to reflect adequately the place of the Atanasoff–Berry Computer (ABC) relative to the Eckert–Mauchly ENIAC in the lineage of the electronic digital computer. (See Alfred E. Brenner’s article, “The Computing Revolution and the Physics Community,” Physics Today, October 1996, page 24 .)

Work on the ABC design by John V. Atanasoff (a University of Wisconsin PhD physics graduate whose adviser was John Van Vleck) began in 1937 at Iowa State University (ISU). It is well established that a breadboard mock-up was completed in 1939 and that a full-scale prototype was being tested by early 1942. The review correctly indicates that, years later, Honeywell initiated a lawsuit claiming that ENIAC patents applied for by Presper Eckert and John Mauchly in 1947, though not issued until 1964 to Sperry Rand, were invalid.

On 19 October 1973, the trial judge entered his opinion, stating that “Eckert and Mauchly did not themselves invent the automatic digital computer, but instead derived that subject matter from one Dr. John Vincent Atanasoff.” ¹ Behind that terse statement is a trial record that exhaustively examines the “prior art” embodied in the ABC and the adoption in either ENIAC or the later EDVAC of many concepts first introduced in the ABC, such as regenerative memory, base-2 calculating, modular construction, and fully electronic computation. ^2,3 (See also Alan R. Mackintosh’s article “The First Electronic Computer,” Physics Today, March 1987, page 25 .)

Evidence introduced at the trial showed that, starting in December 1940, Atanasoff met with Mauchly, briefed him on the ABC design, invited him to Iowa to see the full scale machine under construction (he stayed at Atanasoff’s home) and provided him with free and open access to detailed design features that later appeared in the ENIAC or the EDVAC. Nevertheless, many supporters of ENIAC’s historical primacy still claimed that the court decision was flawed, that the ABC could never operate, and that the ENIAC did not, in fact, depend on the ABC design.

After 1973, Atanasoff began receiving widespread recognition for his accomplishment, including major awards from the Institute of Electrical and Electronics Engineers (IEEE) and the Navy, several honorary doctorates, and, in 1990, the National Medal of Technology presented by President George H. W. Bush.

In 1994, senior engineers at the Department of Energy’s Ames Laboratory put forward the idea that the availability of ABC documentation and old parts could make it possible for them to build a full-scale replica of the ABC that might refute the charge that the ABC could never have operated successfully. A small group of ISU officials, of which I was one, then took on the challenges of project oversight and fundraising.

In late November 1996, the completed (but not yet operational) replica was unveiled in Pittsburgh, Pennsylvania, at “Supercomputing ′96,” an annual joint meeting of the IEEE and the Association for Computing Machinery. The ABC anchored an extensive display of historic supercomputer artifacts in celebration of 50 years of computer development.

By October 1997, all systems were fully operating and the machine was brought to Washington, DC. At the National Press Club, the ABC carried out its first public calculations before computer experts, ISU alumni, and the press. For the next eight months, the ABC toured Iowa, promoting ISU eminence in developing advanced technology. Along the way, some computing runs were videotaped, preserving a visible place in history for Atanasoff’s dream. Ironically, the unattributed adoption of some of the ABC’s concepts apparently provided the only means by which they were incorporated into the mainstream of computer development.