Browse Books

Reviewer: Robert L. Glass

If you need to know something about the Software Engineering Institute's (SEI) Capability Maturity Model (CMM), then you need this book. The big caveat here is that most people who need to know about the CMM have probably already read a great deal of the material that has emerged over the past six or so years from the SEI. They probably have read the book's predecessor, Watts Humphrey's 1989 book [1]. So if you fall into that category, what you want to know about this book is how much of it is new. The answer to that question is “not much.” As the book explains, the new parts are chapter 6, presenting a case study of an organization that has been successful in using the CMM's concepts, and three appendices. Over half of the book is spent on an extremely elaborate discussion of the key practices areas (KPAs) advocated by the CMM, including the goals of each practice, the commitment needed to achieve those goals, the capabilities required to perform the activities, the activities to be performed, the measurement and analysis required to support the activities, and the verification needed to evaluate the implementation of the practice. As if those more than 200 pages on the KPAs are not enough, one of the appendices contains just under 30 pages on an abridged version of the same material. If the case study is the major new contribution of the book, the next question is How good is it__?__ The answer here is “very good.” The case in question is the IBM (now Loral) Space Shuttle Project, whose software achieved a CMM-like level 5 rating in 1988. The project and its use of processes are described in excellent detail. Particularly useful are these lessons learned. In management, empower the technologists. Let responsibility for the process move toward the people who do the work. Decouple the manager from management (as, the book explains, Humphrey earlier recommended). Let quality predominate over schedule in importance; manage error rates, not milestones. Focus responsibility on the team, not the individual. No one individual should be able to cause a failure. Use training and mentoring to raise the level of individual, and therefore team, skills. Using historical knowledge and data, the Shuttle can predict cost to within 10 percent. Only one deadline has been missed in 15 years. In process transfer, specific approaches do not translate well to other projects, but principles do. To transfer principles, transfer managers and technologists (not just documents). In that lesson, there is, of course, an irony. The SEI and the CMM place heavy emphasis on the importance of the use of process to damp out the effects of individuals. Yet, in the final analysis, the transfer of the effects of the CMM to other projects can best be achieved through the use of people. So, the book offers a good case study and a lot of redundancy for those who have previous materials. Should you buy this book__?__ Yes, if you do no<__?__Pub Caret>t know much about the SEI CMM, and you want to learn, or if you know quite a bit about the CMM, but your sources are scattered, and you would like a nice, concise place where you can find all of the information you need. If you meet either of those conditions, you will find in this book a complete, readable, well-organized summary coming straight from the horse's mouth.