Software Architecture: Organizational Principles and Patterns (Software Architecture Series) (Paperback)

Much of the book is devoted to VRAPS. After a few introductory chapters, each principle is defined, explained, and then illuminated with criteria, antipatterns and patterns.

Chapter 8 introduces a case study, based upon a well known Internet company. Allaire's jouney through each principle is discussed, including successful practices and warning signs.

Chapter 9 is added for completeness. It presents a case study about building and implementing a benchmark framework for VRAPS. The authors surveyed many organisations in compiling this book, and their results are published here in a summary form.

The book is rounded off with useful Appendixes. One provides a quick reference principles, criteria, antipatterns and patterns, the other is an index of patterns and antipatterns cross-referenced to principles. Overall I found the book to be well structured and well organised - and not too hard too read.

This is the third book I've read from the Software Architecture Series, and the most practical and useful so far. This book should go far in establishing a basic process for Software Architecture that is both theoretical and practical.

The book is divided into three parts. The first deals with and overview of the VRAPS model and the last deals with a presentation of its application in the context of the situation at Allaire. Both of these sections are fine. It is the middle portion, which attempts to "prove" the validity of their model with reference to object-oriented patterns, which stretches the credulity of the reader. The authors even admit that you can skip the chapters where these patterns are presented. Putting them in only serves to detract from the other meaningful chapters of their work.

Object-oriented analysis was developed to solve abstract problems in reusable code paradigms. Previously unknown to me, a group of eager souls has tried to extend this metaphor into the realm of individual and group psychology, with no attempts to provide an empirical basis for their efforts. Whimsical patterns such as "Antigravity Module," "Drop Pass," and "Loan Shark" are relied upon to produce a catalog of objectivist modes of development behavior.

Why a methodology with no basis in psychology is supposed to answer complex issues of organizational behavior is beyond me. Could anyone conceivably apply theories of database normalization or compiler design and hope that they would help you to manage your employees better? There is a vast body serious material in the business literature which deals with these issues in the proper context, and I see no evidence why adding a layer pseudo-scientific organizational patterns to the analysis can do anything but to add confusion to the problems.

If you want a pop psychology approach to managing behavior in a development environment, a book such as "Dynamics of Software Development," by Jim McCarthy provides a very entertaining and enlightening approach to the problem. If you want an exposure to serious software architecture methods, a tome such as "Software Architecture in Practice," By Len Bass, et al, from the Software Engineering Institute can provide deep insight.

The authors of "Software Architecture" are highly seasoned professionals with impressive experience. I find it difficult to understand why they would take a reasonable approach and burden it with a false sense of analytical rigor based on the patterns literature. They would have had a much better book had they simply focused on the case study of Allaire in the context of the VRAPS model.