June 14, 1999

BOOKS OF THE TIMES

'ENIAC': Creating a Giant Brain, and Not Getting Credit

By CHRISTOPHER LEHMANN-HAUPT

ENIAC The Triumphs and Tragedies of the World's First Computer By Scott McCartney Illustrated. 262 pages. Walker & Company. $23.

ho invented the computer? asks Scott McCartney, a staff writer for The Wall Street Journal, in his diverting new book "ENIAC: The Triumphs and Tragedies of the World's First Computer."

Not Radio Shack, despite the credit given by a child in one of McCartney's anecdotes. It wasn't even John von Neumann, the mathematician and logician who is often called the father of the computer and whose name still labels the basic architecture of the computer.

Instead, McCartney writes, the distinction rightly belongs to John W. Mauchly and Presper Eckert, two men at the University of Pennsylvania who created ENIAC, "the first digital, general-purpose, electronic computer -- the first Giant Brain."

Designed and built in the mid-1940s, ENIAC (for Electronic Numerical Integrator and Computer) was, in McCartney's words, "a bus-size mousetrap of 40 nine-foot-tall cabinets filled with nearly 18,000 vacuum tubes and miles of wiring." He continues: "It was developed as a weapon of war, a machine that could calculate trajectories for World War II artillery guns. Though unwieldy, ENIAC was wired with enough innovation and genius to provide the spark for computer development. The genealogy of the modern computer begins here, with ENIAC, Eckert and Mauchly."

Why is this not more widely known? Why don't the names Eckert and Mauchly resonate like Fulton, McCormick and Edison? The answer to this question is the heart of McCartney's book and provides it with its spice and dramatic tension. First, because ENIAC operated under Army secrecy, it was little known except in computing circles.

Second, complex controversy surrounds who got credit for the machine. As McCartney tells it, while ENIAC was being built a team member met von Neumann, who was interested in crunching numbers for the Manhattan Project, the secret program to design and build the atomic bomb. Von Neumann joined the team, contributed ideas for ENIAC's successor, EDVAC (Electronic Discrete Variable Calculator), and wrote a widely distributed paper on those ideas, "First Draft of a Report on the EDVAC," which included certain of Eckert's and Mauchly's concepts without crediting them. Eckert and Mauchly countered with their own paper, but it lacked von Neumann's eloquence and received little notice. So von Neumann garnered the fame.

Finally, even more complex controversy surrounds who got credit for the very idea of the machine. Very briefly: Eckert and Mauchly failed to file for a patent for ENIAC until June 1947, long after its completion, when they had decided to build a business around their invention. But they proved inept as entrepreneurs and in 1950 had to sell to Remington Rand, the typewriter maker, which five years later was bought by the Sperry Corp. and renamed Sperry Rand.

Sperry, now holding the patent for ENIAC, subsequently signed a cross-licensing agreement with IBM, which by the 1960s had become the leading seller of computers. Challenging this agreement, the Honeywell Corp. in 1967 sued Sperry over the validity of the ENIAC patent. The resulting trial came down to the issue that the Honeywell lawyers unearthed: whether Mauchly had been inspired by John V. Atanasoff, an Iowa State professor who in the early 1940s had been working on a digital calculating machine and whom Mauchly had visited in June 1941. Mauchly proved a poor witness on the stand, and the judge ruled for Honeywell, taking the tack that the ENIAC patent was invalid because it had been applied for too late.

Alison Victoria/ Walker & Company

Scott McCartney Also Today in Books Featured Author: A. S. Byatt First Chapter: 'Against the Tide' Books Forums

McCartney praises the judge's decision, arguing that it was necessary to open up the computer industry to competition. But on the issue of Mauchly vs. Atanasoff, he sides with Mauchly, arguing that whatever "theory and principles" of the computer "had been articulated" before Mauchly went to work on ENIAC, Mauchly had worked with "blinders on" and his solutions were "born of simple practicality." But it is in arguing this point that McCartney's case for ENIAC's originality seems weakest.

Tracing the ancestry of ENIAC, McCartney touches on everything from Charles Babbage's Analytical Engine (1833), which proved impractical to build at the time, down to Howard Aiken's Mark I electromechanical calculator built for IBM in the early 1940s, which unlike Babbage's machine lacked a "conditional branch," the "If ... then statement of computing" known as a subroutine.

McCartney writes that Mauchly claimed never to have heard of Babbage and that he came up with the subroutine idea for ENIAC on his own. But one finds it hard to believe that Mauchly was unaware of computer theories that were in the air at the time, so to speak. Curiously, McCartney makes no mention whatsoever of a principal propagator of those theories, Alan M. Turing, the British mathematician and logician who in the 1930s solved many of the problems of computing with his design of the hypothetical Turing machine.

Still, McCartney makes a persuasive case that ENIAC was the first working computer ever built. Mauchly was heartbroken over never getting proper recognition or profiting from his invention. He chalked up von Neumann's getting credit to the theory, in McCartney's words, "that discoveries are attributed to the person with the biggest reputation at the time."

Eckert, on the other hand, while bitter, "took a more historical view of his fate," McCartney writes, "comforting himself with the knowledge that many inventors didn't receive proper credit until events were digested with wisdom of time."

Maybe the publication of "ENIAC" will prove a part of that process. At least it tells an absorbing story and sheds light on a moment when our world was transformed.

Return to the Books Home Page