The view from the top is not always what we expected from the bottom. The first few chapters of this text were the best I have read in a computer book in a long time, classifying, clarifying, and unifying the various threads of distributed computing in the finest discussion I have ever encountered on the subject. My expectations, however, so buoyed by the first three chapters, were not met by the remaining nine. That is not to say they were badly written, or undeserving of attention. In fact, they covered current frameworks and protocols very nicely, and spent some time describing future frameworks that must evolve from today’s implementations. By themselves, I would normally characterize them as “good.” Yet, somehow, I came to the end of this text, disappointed that it never rose again to the level of its opening.

I still heartily recommend this book. For those who need to understand the history, future, philosophies, and problems of networked programming, this is the book to read. If you lead designers or programmers of networked systems, you need to read this book, and then pass it on to them; they should quote Goff like Lord of the rings fans quote Tolkien. The “Eight Fallacies of Distributed Computing,” attributed by Goff to Peter Deutsch, and referred to in the title, should be taped to the monitor of every distributed program designer, and recited aloud in computer science classes.

The book is an easy, even entertaining, read. Instead of presenting a comparative survey of application programming interfaces (APIs) and protocols, it proceeds from a discussion of the problems to be solved, to the various philosophical postures that have been taken regarding those problems, and, finally, to the frameworks those philosophies have produced. It also discusses how the “laws” or “meta trends” that influence the entire computing industry have shaped, and will continue to shape, network designs and programs. It concentrates on presenting the questions that computer scientists must ask, not on providing oversimplified answers for programmers to type in.

The book has some quirks. A good one, sadly lapsed in contemporary literature of all kinds, is scrupulous attribution. Goff traces every idea to its source, even when the origin is not clear, and ensures that every thinker gets credit for his or her ideas. In addition to raising his personal integrity quotient, Goff has also performed a great service to the reader, by providing a rich resource for additional research and discovery.

A second quirk, which is not good, is a noticeable lack of commas. There are repeated instances of sentences that can be parsed several ways, due to an abnormal and ungrammatical absence of commas. This forces the reader to back up and rescan these awkward spots, usually losing the flow of the discussion in the process.

The final quirk to be mentioned is Goff’s frequent citation of Sun products. If you are a rabid Sun-hater, you will not finish this book. If you are a pro-Sun fanatic, you will rejoice. For the other 98 percent of us, it’s a nonissue that only becomes noticeable due to repetition, much like a mosquito in a dark, quiet bedroom. In his defense, Goff does state his bias up front; after all, he enjoyed working for Sun for a decade, and for six years was a Sun “technology evangelist.” Given his love of Sun platforms, and the fact that this book was released by Sun Microsystems Press, it would not be surprising if it was created on Sun systems. It would, however, be amusing.