The title seems to confuse most people -- this is not a book about using UML to describe a Pattern Language -- it is about a Pattern Language for using UML. Well, about half the book is. The rest is an interesting discursion into the historical evolution of OO patterns, descriptive versus generative patterns, UML, architecture, modelling, and design.
The patterns themselves are fine, ranging over style, substance, domains, products and components. Since many of the patterns were distilled from the Catalysis book, I've moved that up my "to read" pile. However, the patterns as presented here did not impart to me that instant "aha!" feeling that the GoF patterns did. And all the discussion is interesting, in a detached, philosophical way. But the whole thing didn't quite gel for me. In fact, the part I found the most interesting was the description of OO patterns' historical development during the late 1980s and early 1990s.