The Man Who Invented the Computer: The Biography of John Atanasoff, Digital Pioneer [Hardcover]

Jane Smiley's new book is a biography of John Atanasoff, The Man Who Invented the Computer. In this slim volume (219 pages), she attempts to present the life of Atanasoff as well as a number of other contemporaries, describe the machine that he built as well as several other computers, and explore the ENIAC patent trial in 1971. Such an undertaking requires an author with expertise in historical research, physics and electrical engineering, as well as patent law. Unfortunately, Ms. Smiley is an adept writer of fiction.
In addition to a description of Atanasoff's life, the author diffuses her limited effort by including a cast of characters, such as Alan Turing, Konrad Zuse, Thomas Flowers, John von Neumann, John Mauchly and J. Presper Eckert. The bibliography indicates that Ms. Smiley has relied on a remarkably limited number of exclusively secondary or lesser sources. The result is a derivative text that reads more like a college term paper than a serious historical effort.
Ms. Smiley's comprehension of science also appears to be quite limited. She describes quantum mechanics only as "...the science that predicts what happens in systems..." Why mention quantum mechanics at all if this is the best explanation that can be provided? Furthermore, she appears confused at times on the concepts of "analog" and "digital," and seems enamored of the novelty of the binary system, which was actually described more than two thousand years ago, and has been known in its modern form for the last 300 years.
Perhaps the most significant criticism is the lack of clear definition of the terms and concepts that form the basis of the work. The word "computer" has been used over the years to describe both human beings and a broad range of devices from the abacus and slide rule to the Cray Supercomputer. While these uses of the concept may be linked by a "family resemblance," a much more specific definition of the term must be employed if one uses the unique identifier: "the" computer. What are the key characteristics that made the ABC machine the significant break in history that led directly to the modern development of the computer? In fact, the machine that Smiley describes was a modest effort to address a single problem. While there were electronic components, it continued to need human input and control, limiting it to functioning at "human," rather than "electronic" speed. While Smiley states categorically that the machine "worked," this claim is hardly universally accepted. The intermediate storage mechanism was unreliable, and there were problems with both input and output devices. Even the replica made 30 years later fell far short of the identified goal of solving equations with 29 variables.
A critical question remains unaddressed: if the ABC represented a significant advance, a useful machine successfully solving a significant problem, why did no one else see its utility? Why did it sit collecting dust in a basement in Ames until it was eventually scrapped?
Ms. Smiley relies solely on the decision of Judge Larson in 1973 in her attempt to depose ENIAC from its generally acknowledged place as the first computer. She states simply that "...Sperry never appealed the decision, and so they must have accepted it." This is a simplistic and hopelessly naïve viewpoint. Legal decisions result from an adversarial system, where winning, not the ultimate truth, is the established goal. Judge Larson may have been a legal expert, but he knew nothing of the science the competing claims except what was presented in court, and various writers over the years have challenged how much of what was presented he read or understood. A better case than that will have to be presented to replace the first completely electronic, high speed, fully programmable multipurpose computer used successfully to compute firing tables and solve problems related to the hydrogen bomb with a questionably functional dust collector.
In short, this brief effort is no more than a rehash of previously advanced theories, devoid of original research or fresh insight. The advice to "write what you know" is an old cliché, which unfortunately was ignored in this case. Prolific authors occasionally write things that prove embarrassing and publishers sometimes print junk just because of the author's stature. Neither phenomenon is unique to this book, but both are evident. This book rightfully deserves an early trip to the remainder bin.

Jane Smiley, per her reputation as an acclaimed fiction writer, produced a book here that is a gripping tale to read. Unfortunately while she is a great fiction writer, she is absolutely not a historian. This is one of the worst history books I've read in a long time. There are several reasons:
- Her "scoop" is not a scoop at all. The debate of the ABC-vs.-ENIAC has been ongoing for many decades among historians of the computer industry
- Unlike a real historian, who would consider all the available evidence, the vast majority of Smiley's documented sources are "experts" from the university in Iowa -- where her main character of this story worked and studied. It's as if she wrote about the Yankees vs. Red Sox, and all her sources were from New York. Would the 'Sox get a fair shake?
- Smiley has little-to-no comprehension of computers. The ABC was an earlier binary calculating device than ENIAC, but so were dozens of other machines! (The binary issue is one of many mistakes on Smiley's part. She claims that Atanasoff INVENTED the binary machine. In fact, binary was in use for calculating devices decades prior.)
- Another reason the ABC was not a computer is because it had no decision-making capability. It required a human to manually tell it what to do with the results of each step in a math problem.
- Regardless of one's opinion of whether ABC was a "computer" or just a calculator, another problem is the ABC was merely electromachanical, not fully electronic. ABC uses vacuum tubes instead of relays to store its 0s and 1s, but other parts of the computer still uses mechanical equipment. It took ENIAC to be an all-electronic computer (not counting Colossus, because that was a single-purpose machine, vs. the debate here over general-purpose machines.)
- As for the issue of the court case and prior art -- does anyone really believe a judge in the early 1970s understood how computers worked from the 1930s and 1940s? The reasons he gave for deciding in favor of Honeywell had nothing to do with understanding who "invented" the computer. The issues were legal technicalities about how patents are filed.
- Another issue presented by Smiley is, "If the judge was so wrong, why didn't Remington appeal?" By the mid-1970s, with companies like DEC decimating the mainframe business, and with personal microcomputers about to bloom, what would have been the point of appealing ENIAC technology from 1945? Context is key!!!

Anyone interested in an objective view of this debate from real historians should visit their nearest university library and comb through back issues of the "IEEE Annals of the History of Computing" journal.

The parts of the Smiley book concerning American computer developments are superficially researched, riddled with factual errors, and totally biased. It is nothing short of a publishing scandal. I have rated it at one star only because negative numbers were not available.

Smiley's thesis is entirely borrowed from previous writers on one side of the issue. It is that Atanasoff, a brilliant scientist at Iowa State (where Smiley taught for more than a decade), invented "the computer," later called the ABC. Then his ideas were stolen by Mauchly ("a space case") who shared them with Eckert at Penn. Eckert merely "followed through," making sure that Mauchly's designs "were properly executed" during World War II in developing the ENIAC computer for the Army. By contrast, many serious computer historians argue that Atanasoff designed a very limited electronic calculator,while Eckert, who worked closely with Mauchly and others, should be seen as the master engineer of the computer age.

The American portions of Smiley's work (most of the book) rely overwhelmingly on just three second- or third-hand book treatments and an interview with a filmmaker. She directly quotes no documents and offers only one quote from any of the dozens of relevant oral histories--and this one derives from a secondary source. (The only portions of her book that add anything to the record are the oral and written contributions by computer scientist Gustafson.)

The count on Smiley's factual errors is close to two dozen. For example, the two ENIAC leaders met while Eckert was a lab assistant in a prewar crash course in electronics in which physicist Mauchly was a student. Smiley, however, treats them as "lab partners" in a course "in computing theory"--a subject which essentially did not exist in 1941. She says Eckert only had a bachelor's degree by age 27, when he actually had a master's by age 24. She incorrectly states that Mauchly "had run the UNIVAC division" until 1959", when he only ran an applications center within it. She twice cites a statement about the two men's characteristics, once attributing it to Mauchly's widow and once to Eckert's. Beyond factual errors, there are any number of frightfully biased innuendos and interpretations.

Atanasoff was undoubtedly an ingenious man, and this is reflected in his design of the ABC. However, when Smiley adopts for her title, "The Man Who Invented the Computer", it must be difficult for her to allow for any nuance or embarrassing contradictions his story. This is a book that should never have been commissioned (by the Sloan Foundation), written (by novelist Smiley) or published (by Doubleday). Sloppy, sloppy, sloppy.

Peter Eckstein