If you've read Petzold's previous book on WPF "Applications = Code + Markup" the first thing you'll want to know about this book is 'Has Charles made the same mistake and conciously opted to use barely any images or diagrams?'. The good news is no, he hasn't. The book has plenty of screenshots and diagrams that really complement the text.
Petzold has covered the subject in tremendous detail and with a thouroughness I really didn't expect, uncovering mathematical topics I haven't even thought about since I left University. For example, three of the chapters are called "Algorithmic Mesh Geometries", "Matrix Transforms" and "Quaternions" - 3D programming isn't for the mathematically impaired.
As a reference, this book is a *must have* for anyone working with, or about to work with 3D in WPF. On the other hand, if you're looking for a quick overview of the 'art of the possible' with WPF 3D then this probably isn't the book for you. Indeed I think it might have benefited from a longer, gentler introduction to ease you into the subject matter; a 10,000 feet view, if you will.
The prose of the book is very dry making it an almost impossible casual read but, given the subject matter, it would be hard to avoid this problem. So in summary: an OK read for the curious, a great reference and good book overall.
Definitely worth buying if you're using WPF professionally because of the amount of time (and therefore money) it can save compared with just using the standard online documentation of WPF. Has a really good range of examples showing interesting things you can do with WPF and is very readable, conversational and engaging. Like many WPF resources, it's frustrating in having so much material only in XAML and so little in C# - a problem if you need to write an application that's highly dynamic and interactive rather than just using WPF to make an interface look attractive and show ready-made images/meshes. There is a chapter on algorithms to make meshes but that doesn't compensate for the lack of programming-language examples elsewhere. There's rather too much glossing over tricky details and non-obvious syntax where the textbook could really add value, and too much reliance on library functions which are available to download but never explained in the text. There's a lot of basic material on vectors and transformations which you don't need if you do Maths/Physics/Engineering but which might be useful for Computer Scientists or those with a non-technical background. Has a large section on quaternions which is more likely to be useful since this is a more specialised topic.