Reviews for SOA Design Patterns (Prentice Hall Service-Oriented Computing Series from Thomas Erl)

1. A book to state the obvious

   Rated 2 out of 5 stars, February 12th, 2010

   I am (was) a big fan of Erl's SOA series....but this book has been a huge disappointment.
   
   Here are a couple of example of the patterns described in the book:
   
   "Problem: How can unnecessary data validation be avoided?"
   "Solution: A program can be designed to only validate the relevant subset of data and ignore the rest"
   "Problem: How can services interoperate when using different data models?"
   "Solution: A data transformation can be incorporated to convert data between disparate schemas"
   
   These are not patterns. Patterns are a great way to communicate and teach design best practices...
   BUT THIS BOOK IS JUST A COLLETION OF STATEMENTS OF THE OBVIOUS!!
   If you have a minimum real-world SOA experience this book is a waste of time.
   
   The fact that the book has been in progress for a long time is not an excuse.
   Seminal patterns books (Enterprise Integration Patterns, Elements of Reusable OO Design, etc) continue to offer useful insights.
   
   This is well written, but SHALLOW, book. Not relevant.
   There are better ways to spend your time and money.

2. Useful but not a must have

   Rated 3 out of 5 stars, December 12th, 2009

   This book is obviously trying to do for Web Service/SOA what 'Enterprise Integration Patterns' by Hohpe and Woolf did for Messaging Architecture but doesn't quite get there.
   
   Where EIP was so ground breaking was that it combined a new series of diagrams (a la UML) to explain how things connected in each design pattern and contained useful code examples in multiple languages with an associated couple of pages of explanation. SOA Design Patterns has copied the above ideas but the images are 3d icons which makes them difficult to tell apart, with no code examples and fairly ropey case examples rather than actual code.Some of the design patterns are given short shift, especially event driven messaging, but they cover the most important parts of SOA patterns, although what's with everything being prefixed with 'canonical'?
   
   Also, I wonder about the sheer volume of plaudits from industry bigwigs. It starts on the front inside cover and carrys on for 6 pages! It feels as though someone is trying too hard to convince the potential buyer than this book is 'phenomenal', 'excellent', 'impressive' or some other superlative.
   
   Overall it's a useful book rather than a must have.

3. At last, proper enterprise level patterns

   Rated 5 out of 5 stars, September 12th, 2009

   This book is best for those working in large enterprises who have big questions about how to govern their SOA inventory. Please note though, it isn't a programming book, it won't help you write a WSDL, it won't help you select an ESB.
   
   Where it is extremely useful though, is that it clearly gives practical advice and it has helped solve some anti-patterns that were forming in our SOA. Without it and "SOA Principles of Service Design" we would have never been able to achieve the benefits of SOA. Anyone familiar with Thomas Erl's books and articles will know that he focusses on a clear definition of terms for SOA and the patterns re-enforce that.
   
   It's fair to say that the book is aimed at those who 'own' service inventories and architects who have to govern SOA process. Sure, some of the patterns seem quite straight forward and obvious - however, our organisation still managed to get some of it wrong and this book helped crystallise the issues and focus on refactoring to "get it right" and achieve our business goals of flexibility, reuse and cost reduction. Before our architecture team focussed on governing by patterns - we did have renegade projects that were adding services that basically were unusable and short sighted.
   
   If there is a weakness, then I would say that the patterns and case studies remain at pretty high level and whilst we all understand "Canonical Schema pattern", we understand the benefits and the reason why it is a pattern - it is very hard to practically produce a common schema and ensure that it can be governed by people or tools. in fact, we've found that some mainstream ESB products simply won't support some of the patterns.

4. A waste of time for practitioners

   Rated 2 out of 5 stars, March 12nd, 2009

   I bought this book looking forward to some practical help in designing a new SOA framework. I started to read the book and had to skip the first 116 pages to get to the first pattern (do you really need to read about what is a "program" at this level). The actual patterns are described in a very high level language which does not really explain the problem the pattern tries to solve and its solution. Even sentences like: FOR EXAMPLE, are followed by more theoretical talk and no real examples.
   This book might be useful for architects working for HUGE banks but not for practicing architects looking for practical clean solutions.

5. Major Milestone

   Rated 5 out of 5 stars, March 12th, 2009

   SOA Design Pattern is a major milestone for IT. I Imagine it will get the same status as the all time famous Gamma/Helm/Jonson Design book.
   
   Mathias Dietrich, CEO Mobile Net Solutions LTD