What of Atanasoff?

I enjoyed T. A. Heppenheimer’s computer-history article “How Von Neumann Showed the Way” (Fall 1990), but I believe he erred by omitting discussion of the 1973 Sperry Rand-Honeywell patent case over priority for the invention of the digital computer. Judge Earl Larson found that John W. Mauchly and J. Presper Eckert, Jr., “did not themselves first invent the automatic electronic digital computer, but instead derived that subject matter from one Dr. John Vincent Atanasoff.”

The available evidence clearly contradicts Mauchly’s claim that he did not base ENIAC on Atanasoff’s ABC computer. Atanasoff, who was a professor at Iowa State College, wrote to Mauchly about his work and demonstrated his computer for Mauchly in June 1941. Atanasoff’s computer used binary notation, cyclically refreshed capacitor-based memory, vector processing, and an arithmetic processor separate from memory. Although his was a special-purpose computer and was not programmable, ENIAC was not truly programmable either, requiring cables to be patched together. Furthermore, ENIAC was a base-ten machine, a step backward from Atanasoff’s binary computer. The court’s decision represented long-overdue acknowledgment of Atanasoff as the true father of the electronic symbolic computer.

Steven Belknap, M.D.
Maywood, III.

T. A. Heppenheimer replies: While Atanasoff deserves credit, it is far too generous to put him in a class with Eckert and Mauchly and to regard his invention as somehow of a piece with ENIAC. ENIAC had two pathbreaking features: its general-purpose capability and its use of vacuum tubes in large numbers. The second of these was a real advance because it required overcoming the limited reliability of such tubes, to open the way for computers with arbitrarily large power.

Atanasoff’s work, by contrast, deserves to be cited along with Alan Turing’s leading up to Colossus, in Britain, and Howard Aiken’s on Mark I, at Harvard University. Atanasoff’s machine used tubes in limited numbers and had no pretensions to generality. It could do one thing only: apply a particular mathematical procedure called Gauss’s method to solve a particular set of algebraic expressions known as a system of first-order equations.

Colossus used some twenty-four hundred tubes, compared with ENIAC’s almost eighteen thousand, and it, too, could do only one thing: crack German codes. Mark I, by contrast, could serve as a general-purpose machine, but the reliability problem forced Aiken to use telephone relays instead of tubes, and they were far slower.

What of Judge Earl Larson’s 1973 decision and the subsequent acclaim that Atanasoff’s advocates have won for him? Ironically, even from my point of view I can see little in ENlAC that could support a significant patent. The concept of a general-purpose digital computer already existed, in the Mark I. And the Eckert-Mauchly method for improving the reliability of vacuum tubes was too specialized to carry over to transistors and integrated circuits, which are inherently highly reliable. Nevertheless, Eckert and Mauchly built a recognizable precursor of the computer. Atanasoff, Aiken, and Turing made important contributions to this invention, but ENIAC was the real thing.

Chemical Creators

I am writing to compliment Drs. Ralph Landau and Nathan Rosenberg for “America’s High-Tech Triumph” in your Fall 1990 issue. As a practicing chemical engineer, I found the article an excellent history of my profession and very enjoyable to read.

The article correctly reports that the profession of chemical engineering was born at MIT. One might also get the impression from the article that the profession grew to maturity at MIT without ever straying from its roots; the work of countless academicians and professionals at other institutions is overlooked. A very incomplete list of universities with excellent chemical engineering departments includes Oregon State University, Yale, New York University, the Universities of Delaware, Michigan, and Wisconsin, and my own alma mater, Brooklyn Polytech (now Polytechnic University). The people at these places have made contributions of such magnitude that without them none of us as individual engineers, nor the profession as a whole, could have achieved the level of technology so well documented and applauded by Landau and Rosenberg.

Norman Ostroff, P.E.
Stamford, Conn.

Prior To Pong

Frederic D. Schwarz’s article “The Patriarch of Pong” (“Postfix,” Fall 1990) identifies William Higinbotham as the inventor of the video game, in 1958. Higinbotham’s was not the only video game of that era, however. As a student at MIT in the early 1960s, I played Space War, an early and surprisingly sophisticated precursor to Space Invaders, on the Digital PDP-1. The game was enjoyed by many students, but the key to commercialization—connecting a coin box to the computer—would not be found until the next decade.

Albert H. Teich
Director, Directorate for Science
and Policy Programs
American Association for the
Advancement of Science
Washington, D.C.