[chrismear]

Maybe this is the issue -- I've seen these things explained the other way around.

My experience is in Ruby (specifically Rails), with Test::Unit (usually used to espouse TDD), and RSpec (recommended for BDD-style development).

With TDD, I follow the pattern of using real objects for almost everything (only mocking out API wrappers, etc.), setting some initial values, poking a method, and then testing that some final values are as they should be.

With BDD, I end up mocking and stubbing almost everything, and then writing a test that checks that my object under test calls the right methods on other objects.

I like the BDD school for its focus on testing one class in isolation, but often when writing BDD-style tests I just end up echoing the (eventual) implementation in test-form, line by line. It's entirely possible I've got utterly the wrong end of the stick, but I can't shake the feeling that it seems a bit redundant at best, and of little added value at worst.