Brian's analysis is awesome.  I'm not convinced that I agree with his biases of preferred mechanisms - but I do come from a developer who tests perspective.  If I were asking a non-developer to write the tests, then I think I would agree with his biases.  Either way - a great analysis.