This sounds like a very interesting book. The one small concern I might have (I have only read the review) is that the sample problem might not be a good case for some of the more complex techniques, such as introspection and metaprogramming. I think we have all seen code that seems to have been designed primarily to showcase all of its author's knowledge regardless of whether it helps, and I would not want the book to be misunderstood as an endorsement of that practice.