Contrary to what book authors want you to believe, XSLT is not easy to use. Someone said it's "not fit for human consumption". And Adam Bosworth has said "This is the paradox: XML was chosen in part because humans could read and write it, unlike the highly efficient babel of binary formats that preceded it. Yet languages encoded as XML grammars and used for manipulating XML can only really be read and written by programs (and a few very smart people)."
G Ken Holman is a very smart person. He's written a book that is a very strong investigation of the theory behind XSLT and he works very hard to try to bend the mind of the reader to a new way of looking at processing XML. Many readers will be used to procedural text preprocessing, like I was, and this book sets itself apart from the rest with this very authoritative and thorough background. Ken frequently addresses higher-level concepts like how to manage stylesheets and how they fit in context with other XML and web technologies. Dont' get stuck trying to use XSLT for something it wasn't designed for!
I looked at 5 other XSLT books and in this one the examples were more varied and re-world and explained in more detail. But a million examples will do no good if you make one modification, it breaks and you don't know why. This book's primary contribution is answering the 'why'.
I didn't have to play with xslt for long to realize that you can get stuck for a long time on a tiny little bug that isn't evident unless you have a full explanation of the standard. XSLT and XPath syntax drives me crazy and it can be very hard to 'see' what is going to happen when the transform is run. There's probably no fix for that. This book is properly focused on trying to fill in the holes in your understanding so you can code effectively.