Reviewed: November 2012
What a difference there is between this book and the subject of my book review, Practical Software Measurement . I suppose that in a way that makes them rather complementary. This book is very much concerned with the practical side of measurement and statistical analysis.
The first two chapters cover background and planning what to measure. The core of the book then covers collecting data, analyzing it and using it for process improvement. The tone and language is similar to that of Deming, talking about the process being under statistical control (or not). If the process is not under statistical control, then it’s because there are assignable causes. Once you’ve identified and eliminated assignable causes, then the remaining variations are inherent in the process, and you can set about reducing that variation.
The final two chapters give advice on how to improve your process based on the measurements and analysis and some practical tips covering getting started and a FAQ. The appendices mostly cover the statistical methods that are required for the analysis described in the book.
I felt that the best part of the book was the advice on how much data to collect and how soon to start using the analyses. The main thing that seemed to be lacking was real-world examples. These methods seem to be commonplace in manufacturing industry but rare in software development. There are some examples in the text, but too short to give any real feeling as to how much benefit can be had from applying measurement in this way.