Search in Book Reviews

There have always been many high profile software project failures - and there are examples today. So why has there been no apparent improvement? A difficult question with no simple answer. Some would argue that, as an industry, we have not properly implemented mechanisms for measuring the software process and that is where this book comes in, by applying statistical process control for software process improvement.

The authors cover what the software process is, planning for measurement, collecting data, analysing process behaviour, producing process behaviour charts and three paths to process improvement. The material is clearly explained but does include some statistical formulae and these are a barrier to understanding what is being achieved. This is the type of book that requires being read more than once to ensure that the concepts are fully understood and could be applied. The essence of this book is defining a mechanism to obtain objective information (as opposed to data) that can be used to control and improve process. Therefore the audience for this book is aimed at team leaders/managers rather than the developers themselves. For many leaders/managers it is very difficult to understand what is going on in the trenches and trust plays a significant part in controlling software development. This book details an alternative, however, applying that may not be very straightforward.

To address that concern of how to get started the book devotes a section to "Getting Started" and also some frequently asked questions. Probably the best advice contained is to get a software package that can perform the statistical calculations and produce the appropriate charts! This is not the easiest book to understand, especially for those not statistically oriented, but the subject is so import that it deserves some attention.