Well, you should read his recent posts on why Frederic Brooks's "No Silver Bullet" was wrong:

http://rebelscience.blogspot.com/2010/02/reliable-software-p...