Perroud and Inversini give us a book that discusses issues of “recurring architecture problems” in information technology systems, with the emphasis upon major enterprises. The subject is a fascinating one, well deserving of attention. The authors have done their background work well; they clearly show quite a level of expertise in their subject.

There is a small issue with the quality of the use of the English language, which is obviously not their first language. There are some idioms that seem to be unnaturally translated, leaving the reader to puzzle out what the authors are trying to say. Luckily, this does not happen frequently, but it does slow the pace of reading. This is certainly not a major criticism, but it may be an issue for some, nonetheless.

The authors use the metaphor of building architecture quite a lot, even talking about “IT-bricks.” They overuse this metaphor somewhat. They failed to convince me that “bricks” are really what we are most looking for in major information systems. (It sounds like the old arguments concerning modularity, but maybe I am just showing my age by remembering those.).

The book is divided into sections about theory, catalogues, the use of patterns (specific, infrastructure, and general), and pattern symbolism. It is well designed. Generally speaking, I like it, but, problematically, I am unsure of its source. The authors speak of enterprise systems, usually for large enterprises. They wish to speak of recurring patterns (see the subtitle) across such systems. However, such major systems usually do not recur within a specific enterprise, but they may recur across several enterprises. In my experience, that type of information is not easily shared between rival companies, and someone who shares it may very well find himself both fired from his job as a traitor and unable to find similar jobs with a competitor. My experience is that major information systems are critical competitive aspects between business rivals. Am I so out of date that I am unaware of forums for sharing such information, across multiple enterprises, with the agreement of enterprise management? Frankly, I somewhat doubt that. On that point, the authors leave me unconvinced. In other words, while the book is well written (aside from that noncritical English language issue mentioned before) and well constructed, the central assumption of the book may be less implementable than the authors appear to believe. If such is the case, then the book is good, but may have been released a bit prior to its optimal timing. Of course, more power to the authors if that is true!