Painstakingly, in a literal sense, read from cover to cover without learning much about semantic description and search (more pedestrian XML technologies, eg, XPath were covered well). Some of it, eg, on Topic Maps, is impenetrable. Very light on interesting and compelling usage and how-to of the more ambitious, semantic technologies that are the reason most would buy a book of this title.
And so, unfortunately, I agree with the negative assessments already given here: little practical information for implementers and on the contrary, the considerable time spent in attempts to decipher will not be justified, in my experience, with their pay off in knowledge that is useful or memorable.
To be fair, part of the problem, from what I gather by its absense in the book, is that the W3C semantic web technologies are not even attempting to solve any part of the ultimate problem of semantic analysis: natural language understanding. Instead the highest goal in this presentation is the /manual/ cataloging of /whole/ documents (and emails, customer questions, etc).