The good people from Packt Publishing asked me to review another one of their SOA books: Service Oriented Architecture: An Integration Blueprint . The book has arrived, but before I start reading it I wanted to capture some expectations of what I hope to find in and learn from it. These expectations (highlighted in green – the color of hope) will also serve as evaluation criteria when I will be writing the actual review.
In order to be methodical let us start with the intended audience: I have written and spoken on SOA and other subjects numerous times and always tried to tailor the content, structure, language and style to the audience I was planning to address. The book is aimed quite widely by both the role: Architect to Manager and by level: on one hand the book requires comprehensive understanding of SOA, while at the same time promising that it will be useful to the uninitiated. So I expect that the core audience is senior technologists skilled in the art of SOA and Business Integration. I happen to belong to that demographic, so I can expect to learn something interesting and new from the book.
The book vows to provide a theoretical introduction to SOA using the Trivadis Integration Architecture Blueprint. This made me somewhat worried – the first word meant nothing to me and it sounded as a methodology. I am weary of methodologies in general and especially the ones I never heard of. I believe methodologies and processes are like Constitutions: one has to be of truly remarkable abilities to come with one that works when practiced by the masses. And even then it might be of questionable value. Ivar Jacobson one of those few who actually came up with a methodology which stood the test of time recently wrote a manifesto Enough Process – Let’s do Practices. After some googling I found out that Trivadis is a small Swiss IT consulting company working primarily in the German-speaking part of Europe. If there is anything I find more suspicious than obscure methodologies it would be private corporate brand methodologies. I am speaking from experience – I once found myself a custodian of such a methodology called RQ which I think stood for Repeatable Quality. As far as private methodologies go, it was not the worst specimen – it was not written by a couple of junior consultants whilst being on the bench, but actually came from the assembly line of a respected methodology-building company at a cost of nearly one million dollars. And to me it still looked like a slightly tweaked RUP with some added domain-specific terminology and document templates. I am hoping that Trivadis Blueprint would not end up being one of those methodologies.
Here is the list of what I am expecting to find in this book based on the information available from the publisher:
- A systematic, comprehensive and cohesive conceptual model of the Enterprise Integration Domain.
- An unbiased overview of relevant technologies and implementation alternatives.
- Some battle-proven best practices and implementation guidelines.
- Some war stories and lessons learned from the field.
- A picture of how SOA relates to integration and where are the impedance mismatches between the two.
- Contain an original idea – something that is a) non-obvious and b) cannot be traced to an earlier source after 15 minutes of googling.
- A peek beyond the horizon at the things to come.
I am starting my reading tonight and will report on how the book measured up against these expectations.