Facts and Fallacies of Software Engineering
by Robert L. Glass

ISBN-13: 978-0321174253
Publisher: Addison-Wesley Professional
Pages: 224

I’ll go directly to the point. Robert Glass is a great writer and you really want to read this book. Not exactly that it presents anything new; quite contrary, Robert Glass has collected 55 facts and 10 fallacies of software development. Particularly the facts are based on observations that people in the field all know or once knew, yet frequently forget. For example, software estimates are typically done at the wrong time, often by the wrong people, and are rarely corrected as the project proceeds (facts 9-11). If you ever been asked for or were evolved in software estimates you know that software estimates are particularly poor and, very often, based more on wishes and deadlines than actual realistic targets. Yet people live and die by software estimates anyway (fact 12) and get tremendously concerned as projects don’t meet them! Most of the book is indeed about an industry that refuses to learn.

The material is clearly presented with insightful discussions and comprehensive references on each fact. A most positive aspect is the table of contents; it’s actually readable as it stands and gives a quick overview of the facts and fallacies, just like a distillate of the whole book.

The first part covering the facts contains little controversy. Anyone with real-life experience from the software industry will agree with most of it, although there are huge discrepancies about the cures to the problems (“prescriptive”, “CMMi”, “no, agile”, “XML”, “frameworks”, “draw-some-diagrams-and-outsource-the-pressing-of-the-button-that-generates-the-whole-application tools”, you get my point). The fallacies-part of the book changes this. Open source advocates will probably not like to see the classic mantra “given enough eyeballs, all bugs are shallow” presented as a fallacy! No matter how much I like open source and how valuable it is, in the long run, making claims that aren’t possible to back-up will backfire. It’s important to get this right; Robert Glass does not suggest that open source is either worse or better than its counterpart. It’s just that there is simply no evidence that the claim of “enough eyeballs” is true. Robert’s treatment of the topic goes deeper and is very thought provoking. I really do recommend that you get a copy of this book and study it yourself. It’s praiseworthily short and tremendously interesting.

Reviewed March 2007