Object-Oriented Software Engineering: Using UML, Patterns and Java (draft 1999)

The only useful chapter in this entire book is the last one, where the book finally asks the question "How much process/modeling/management/etc. is enough?" for a given project. A number of methodologies are presented, that combine just a few of the techniques covered in the other 15 chapters, and discussed in the context of different types and sizes of project. The rest of the book, comprising 15 chapters, is structured as a catalog of modeling and analysis techniques but they're presented as *the* way to develop software, according to the authors. It's only until the end that those techniques are placed in a context where they make any sort of sense. The book also uses UML to model the activities and techniques in each chapter. This is cute, but not especially helpful and occasionally confusing. In fact, this usage serves as a helpful counter-example to the idea, espoused in the book, of UML as a catch-all modeling tool. In lieu of this book, computer science students would be well served to study a range of methodologies and projects to which they were applied, and then to dig deeper into the techniques that make up those methodologies, according to the interest and needs of each student, in other dedicated references.