I've read the whole book once, and several sections a couple of times; and my impression is that - although good - it doesn't really seem to know what its target readership is. At times, it's very academic, going into the history of CSS and HTML development, explaining the theory of document analysis, and drawing comparisons between print and on-screen typography. The theoretical parts are all very interesting and well worth reading, but I found it a bit odd that a 10 chapter book took until chapter 5 before actually getting on to how to write CSS rules.
The chapters on typography (more precisely, text styling) and boxes (the CSS concept of how things are arranged on screen) are very good, and probably worth the price of the book alone. In spite of the academic tone of the opening chapters, the book deliberately roots itself in the real world. It states from the outset that it doesn't intend to cover every rule in CSS, only ones that work in today's browsers. It highlights the problems with N4 - and offers suitable workrounds. It also emphasizes that there's no single answer to design issues - it's a matter of experimentation until you achieve what looks right for you in a variety of browsers.
Because the book is relatively short (280 pages) and the chapters are well laid out, it's easy to find things you want to go back to. Surprisingly, though, there's no quick reference list of the rules covered, so it's impossible to see at a glance the areas the book actually covers. One striking omission is in the section on the shorthand version of font. It says "it first takes values for each of font-style, font-variant, and font-weight", but nowhere in the book could I find a definition of font-variant.
It's a pity that what could have been a very good book has got such flaws. Newcomers to CSS would learn a great deal from it, but may be put off by the approach.