Browse Books

Reviewer: Leo G. Egan

This book describes a necessary step in the software development process. Too many software managers and software companies do not know how to estimate software costs properly. A careful reading of this book will enlighten all to a disciplined approach to estimating the costs of software. Unfortunately, though, it requires a great affinity for mathematics and comfort with statistical analyses. There is a strong emphasis on mathematical analysis, and the reader must have a good foundation in statistics to conquer the material. The book is exciting and fairly easy to follow if you do not get bogged down in the math. However, the reader may have difficulty following the mathematical derivations because Londeix does not include the intermediate steps. Therefore, the reader must accept the author's conclusions, which appear logically correct, at face value. Londeix's approach would be easier to use if a more complete explanation of the method were provided. What is needed in the software development world is a unique and fairly easy way to identify software values and to be able to crank these values into an equation (or software program) and come out with an answer. This book frequently comes across like a treatise on mathematics rather than a treatise on how to estimate software costs. But the cost estimation equation would serve its purpose well if it were easier to understand and use. The concept of a product plan, as derived by the author, is unique. However, it is not clear what the reader is supposed to do with this concept. A concrete proposal for the use of the product plan is essential. I like the idea of a consulting estimater developing an estimate. However, this estimate should reflect the actual workplace requirements. Estimates should influence the scope of the project rather than the reverse. The book includes three project scales apparently put forth by Putnam. These are not universal standards accepted by industry. However, the author should have ended up with equations for each scale project. Management is typically responsible for estimating costs, even though others gather the data. However, no matter who calculates the estimates or collects the data, software estimates need to have management understanding and approval. Therefore, I would recommend this book to executives who need to understand how software estimates are generated. I would also recommend this book for university use. It would be excellent for an economics curriculum at the senior undergraduate or master's level. A sequel that ties the software development life cycle process to cost estimates and eventually to cost control is needed.