Browse Books

Reviewer: James W. Moore

Of all the books I have reviewed, this is, perhaps, the most requested by my colleagues. Each of the authors is respected in the field and can claim significant achievements spanning the broad variety of topics covered in the book. Expectations are high. The book is intended for software practitioners and managers who want to improve their development processes by incorporating practices for systematic software reuse. The authors regard object technology and software reuse as vitally linked; they believe that success in one requires success in the other. The claim is large: “systematic, large-scale reuse, coupled with object technology, is the only way to radically improve the process of software development.” The scope is broad: the book claims to provide a “pragmatic framework for success” dealing with “business, process, architecture, and organization issues.” The goal is grand: “building a Reuse-driven Software Engineering Business (RSEB).” The book is divided into four parts: an introduction to principles of systematic reuse; a discussion of architectural style, which is largely devoted to an explanation of the Unified Modeling Language (UML) as applied to reuse; a description of software and business processes intended to exploit object technology in the pursuit of large-scale reuse; and a discussion of organizing a reuse business, which provides advice on exploiting the principles of reuse. Annexes provide a glossary, an annotated bibliography, a list of references, and a summary of UML. The text includes many examples and figures, most of which apply UML. To implement the RSEB, the authors suggest a staged transition approach requiring “simultaneous changes in…business, people, process, organization, architecture, tools and technology.” They encourage the use of pilot projects but emphasize that “systematic reuse cannot be adopted piecemeal.” In comparison with other approaches, they place increased emphasis on architecture and they apply object-oriented methods to business process engineering as well as to software engineering. In judging such an ambitious effort by such capable authors, a careful choice of yardstick is required. For example, I might question whether it is presumptuous for software engineers to explain how to restructure a business—but surely no sensible businessperson would take action based solely on such advice. Instead, I have to admire the authors' courage in pursuing technical conclusions beyond their usual boundaries. So, my comments deal with issues well within the presumed competency of the authors. First is the validity of the advice. There have been few successful business-wide applications of software reuse. The authors' successful results in their organizations may not apply equally to other businesses. They should be viewed as examples rather than prescriptions. Second, one should be respectfully skeptical when any single tool—in this case, object-oriented technology—seems to be the solution to so many problems. Solutions to complex business problems typically have real-world messiness rather than technological purity. Third, the discussion of UML is difficult, and some of the language's characteristics seem to be arbitrary rather than useful. I cannot judge whether this is a problem of the language or of the presentation, but the current popularity of the language suggests the latter. Finally, the writing is dry. None of these criticisms diminishes the value of the text. Three experienced practitioners pursue a promising technology in approaching a difficult problem that inherently extends beyond technological considerations. It is to be expected that the technology will be stretched to its limits of applicability—and perhaps beyond. With most important technologies, we cannot find the boundary until we cross it.