Browse Books

Reviewer: Andres Silva

Leffingwell and Widrig have been professionally involved in requirements management (RM) for many years. Their involvement is not only related to the development of software-intensive products and services, where they have extensive experience, but also in the development of RM tools (RequisitePro). In this second edition of their book, which is more focused on use cases than the previous edition [1], the authors communicate their accumulated experience and wisdom to the reader in a direct and sometimes amusing fashion. Throughout the book, there is a feeling that it was written from the trenches, making it one of the most enjoyable and readable books on the subject, despite its length. The book is structured into a series of so-called team skills, defined by the authors as those "skills that are necessary for a modern software team to successfully address the requirements challenge." After an introduction (chapters 1 through 4), the first team skill deals with analyzing the problem (chapters 5 through 7). The other team skills covered are about understanding user and stakeholder needs (chapters 8 through 13), defining the system (chapters 14 through 17), managing scope (chapters 18 and 19), refining the system definition (chapters 20 through 24), and building the right system (chapters 25 through 29). At the end of the book, there are two extra chapters covering agile requirements methods and a prescription for requirements management. The six appendices provide a complete example (a home-light automation system), templates for relevant documents, and some notes on requirements management in the rational unified process, the Software Engineering Institute capability maturity model (SEI-CMM), and ISO 9000:2000. The authors did not include a glossary of terms, however. It is almost a tradition that every book on RM defines, or redefines, the terms used in it; this book is no exception, and the authors make their contribution to the terminological confusion that, unfortunately, is so common in the field. However, their proposed approach, based on a simple "needs-features-requirements" trilogy, is more appealing and user-oriented than other, sometimes convoluted approaches used by other authors. Actually, I used this approach in teaching requirements engineering, and the students felt very comfortable with it. It is also becoming a tradition in requirements books to justify the need for a good requirements process on the basis of data published by the Standish Group in its famous Chaos Report. However, the authors insist on the Chaos Report from 1994 [2], now made obsolete by more recent Chaos Reports that, encouragingly, have revealed small improvements in software and requirements industrial practice. The problem with this book is that it is quite irregular. Some chapters contribute almost nothing to the reader, in contrast with other very good ones. For instance, there are two chapters devoted to business systems and systems engineering. The chapter on business systems ("Business Modeling") is very poor, and does not even scratch the surface of the subject. On the other hand, the chapter on systems engineering ("Systems Engineering of Software-Intensive Systems") is wonderful, very complete, and quite deep, presenting a good analysis of the challenges that those systems pose, particularly when they are developed in traditional nonsoftware industries. Also, the book provides a superficial treatment of the relationship between use cases and how they are implemented (chapter 25, "From Use Cases to Implementation"), which is in contrast to the deep, useful, and interesting treatment of product management (chapter 17), a topic unfortunately not covered by other RM books. Overall, this is an extensive tutorial on RM for software-related systems. The authors did not include exercises or any other supporting materials so, as a standalone book, it cannot be used to teach this subject. Also, use cases (the topic of the subtitle) are not treated in depth, and readers interested in this topic should look elsewhere to find more information [3]. It is worth noting, however, that the intended audience of the book is not an academic one; it is written for practitioners and developers. From this point of view, the authors deliver what they promise. Online Computing Reviews Service