The history of the first Monte Carlo computer simulations was what got me interested in this book in the first place. Once I started reading it, I discovered how little I knew of the Electronic Numerical Integrator and Computer (ENIAC) besides a few bits of trivia, and how interesting its history was.

ENIAC in action presents the development of the first electronic computer in the context of other computing devices of that era and their limitations. Existing analog differential analyzers and electromechanical calculators based on relay technology were complicated, difficult to operate, and slow. The most important difference between the proposed design for what was to become ENIAC and earlier computing machines was its use of vacuum tubes for its processing and storage circuits. This would enable computing at “electronic speed,” promising to clear the growing backlog of calculations to produce firing tables required by the artillery.

The book describes in detail the design process behind various components of ENIAC, the challenges of building the hardware in the face of wartime shortages, the development of the programming method and notation, and creating and running the first programs (back then known as “set-ups” instead).

The original ENIAC’s architecture was different from modern stored-program computers: the programs were set up on it by changing how its different components were wired together and in effect creating every time a new special-purpose computer. It was fast, able to carry out its computations in parallel but limiting the logical structure of the programs (the number of loops, conditional instructions, and so on), and difficult to program and troubleshoot.

Perhaps the most interesting parts of the book are the chapters describing the lengthy process of converting ENIAC to the modern stored-program architecture, where the connections between various components remained fixed, implementing a newly invented set of basic machine instructions, and the programs were entered as sequences of numbers into a form of read-only memory (known as “function tables” in ENIAC lingo). This new approach made programming much easier, lifted the limitations on the logical structure of the programs the “new” ENIAC was able to run, and enabled it to run a more advanced version of Monte Carlo simulation than initially intended. The drawback was that, in the new configuration, ENIAC became a serial computer and its programs ran much slower.

Far from focusing only on the technical aspects of creating and operating ENIAC, the book also devotes much space to the people behind the project: its inventors; the women recruited from human computers, who became its first programmers and operators; and scientists who were its first users. Equally interesting is the account of the contribution of John von Neumann to the stored-program reincarnation of the original ENIAC.

I really enjoyed this book, not just for its wealth of technical details on ENIAC, but also for how it presented the development of ENIAC as a chain of events: causes and effects set in the context of wartime needs.

More reviews about this item: Amazon