Browse Books

Reviewer: Pavol Navrat

Shaw and Garlan have written one of the first books devoted to software architecture. Both have been pioneers in the field for many years, as reflected in this volume, which integrates the results of several years of research. Much of the material in the book is derived from the authors' previously published work, yet the book is not merely a collection of previously published papers. A foreword by Barry Boehm stresses the importance of software architecture as the most important milestone in the software life cycle. Chapter 1 introduces software architecture as an emerging engineering discipline of software; it is based on a well-known 1990 paper by Shaw [1]. Chapter 2 presents several typical architectural styles: pipes and filters, data abstraction and object-oriented organization, event-based, implicit invocation, layered systems, repositories, interpreters, process-control paradigms, and heterogeneous architectures. In chapter 3, five case studies (keyword in context, instrumentation software, mobile robotics, cruise control, and examples of heterogeneous architectures) are elaborated with alternative styles. Chapter 4 is devoted to shared information systems. The fifth chapter, on architectural design guidance, is based on T. G. Lane's doctoral research [2]. Chapter 6, on formal models and specifications, illustrates approaches to formalizing software architecture using the Z specification language. Chapter 7, on linguistic issues, discusses architecture description languages, covering their requirements, syntax, and semantics. Chapter 8 presents tools for architectural design by depicting three research systems. Chapter 9 describes a course that uses the text. I greatly appreciate the authors' role in the development of the discipline. In fact, I admit to having hoped they would attempt in this volume the first truly systematic treatment of the discipline. The book's style falls somewhere between a state-of-the-art reading and a monograph. It covers much of the state of the art, despite relying mostly on the authors' previously published works, because the authors are leading figures in the field, and their work has strongly influenced its development. The contents and the methodology do not suffice for a monograph, however, mainly because the discipline is not mature enough. To be fair, the authors do not state any such aspiration. On the one hand, the book reflects the immaturity of the discipline of software architecture, and on the other, it is an important step toward its maturity. In a mature discipline, a commonly adopted architecture description language will be available, and a monograph will generally use it from the beginning in describing the main architectural styles. Here, the styles are presented informally, and more formal descriptive methods are discussed only much later in the book. The formalisms are still topics of research, as is almost everything in the discipline. We do not yet have a set of standard methods. We do not yet have hard results. We hope to see the open problems in and the motivation for the emerging field of software architecture. In this well-written and highly readable book, Shaw and Garlan offer an excellent overview of what they have accomplished in this field and their perspectives on it. In addition, they include a comprehensive bibliography.