The approach the author takes is to assume some knowledge of UML and then to cycle through different topics looking at the relevant constructs and how they fit together. While there's a knowledge of UML 1.x assumed, the book does not really look at it in relation to UML 2.0. For instance there are no 'this is new' or 'this has changed' sections. What's more there's an implicit assumption that the reader understands object oriented development (which goes without saying if the reader is already familiar with UML 1.x).
Read the rest of the review at TechBookReport"