The lead author of this book is one of the triumvirate who developed the original GML, the ancestor of SGML, XML and HTML. Actually after the introductory chapters, most of this book is written by 'contributing experts' coupled with sponsors. A little thought will tell you that this is not an independent book but a conglomeration of material from interested parties. Let me give you an example
Chapter 31 'Converting renditions to abstractions' starts with three pages about the general concept before focusing on a very specific example of using avenue.quark to convert a QuarkXpress document to an XML one. Unless you have some familiarity with the QuarkXpress, much of this is likely to pass you by. In addition, why pick on avenue.quark? There are other routes to the same objective. Of course it is no accident that Quark, Inc. are sponsors of this chapter.
I have no problem with sponsored books (rather like commercial TV) but I think that charging $49.99 for a book in which the sponsors have so much financial interest (I do not know what sponsorship included) seems a bit steep.
Yes there is a lot in this book. Not least, seeing how commercial organisations are building on XML is informative. I think that for that reason alone this book is worth some time spent browsing it. I do not think you would want to read it cover to cover. Remember that while conferences often add value with vendor presentations you pay to listen to the non-commercial presentations. In the same way, I think you are being asked to pay in this book for items that should be free. On that measure this book should cost about $15. At that price it would be good value.