what he says doesn't work does, what he says does doesn't

Gregor Hohpe should have read past the first 100 pages. This book is good on theory, poor on practice (does that remind you of any other SOA book?).
The examples Havey provides of "non-trivial" systems in the back are, in fact, quite trivial. What's worse is that when he ventures into the territory of "advanced" features, he gets lost. For example, on p.270, he provides an eventHandlers section, but comments it out saying that it doesn't work. I was able to get it to work as written with just a minor tweak, but he slags off the vendor instead (p.284) and proposes an awkward hack for a workaround (p.277). Then, on p.308, he presents us with a piece of parallelism that depends for its success on the use of a correlationSet. This is supposed to be clever, but is, in fact, just poor programming practice. Not only that, but it doesn't work! It can't possibly, not the way it's written. He just sent it off to the publisher without testing. We're not talking about simple syntax errors here... this is a fundamental conceptual flaw in what he's proposing. Pretty basic stuff for him to be stubbing his toe on.

Other Reviews

• Application Architecture relation
• The book is a bit of a misnomer - it should be essential business process modelling for technicians
• Good for understanding the technology and standards
• Best Practice
• what he says doesn't work does, what he says does doesn't