What I like about this book is that the author effectively acknowledges the shortcomings of traditional project management and does not harp on its dogma. The material on adaptive project management offers a way out for those projects where the method of implementation is so unknown that traditional full-court project management practices will only hurt and not help. Much of the information on adaptive and extreme project management is similar to the "extreme programming" methodology, but specifically targeted towards a project manager.
I think this book is an excellent purchase. It is well written and succinct (unlike the 1000 pages of bullet-points from a competing author). Other reasons to purchase the book are its reasonable price and the fact that you get a trial version of Microsoft Project 2002 with it.
I wish the author had gone into greater depth about earned value. I have found earned value difficult to do without good support from the project management tool (calculating BCWS for example is tedious to do by hand); every time I tried to do earned value for a real project using MP98 or MP2000 the tool has crashed on me.
In summary I believe that this is an excellent book, and should be read by every project manager.
To me the value of the third edition lies in Robert Wysocki's recognition of projects where either the goal is clear but the methods aren't, or where even the goal is unclear. These types of projects seem to predominate in IT. This book is worth purchasing for the insight that the authors bring into the non-traditional non-textbook real-life projects.
The prose is clearly written and reads very tightly. Contrast this to the random collection of thoughts in Kerzner's book. This book is written for the practitioner, though someone taking a college course in project management would also benefit.
Excellent book, at an excellent price. You can't go wrong.