Well I finished the book Test-Driven Development in Microsoft .NET and I still feel less then satisfied. My problem with TDD stems from the test first then code idea. I like the idea of being able to test the public interface to a class, assembly etc. This can really be useful in a team development environment. An example of this happend to a co-worker today. Someone changed what was returned by a property. Not sure if anyone figured out why, but a simple NUnit test would have discovered this, and at a minimum made the developer that made the change think about it for a moment.
I try to keep an open mind regarding things like TDD. However there are several of the XP ideas that I really have a hard time with. This book highlighted a couple of the problems that I have. Every other page was spent refactoring code and / or tests. Now I'm all for refactoring code when it is needed, but most of the refactoring that was done in the examples would not have been needed if there was a little, not a lot, but a little up front thinking about the application. Spending all you time refactoring does not make Joe Developer more productive. He may be writing a ton more code, but if he is constantly refactoring the same 3 classes he is not making any progress.
I'm not going to sit here and say that TDD and XP in general are stupid and should never be done. What I will say is that you shouldn't do them just because you read about it in a newgroup or someone's blog. Keep in mind that C3 was a failure from the customer standpoint, and that is the one that counts.