[johnwatson11218]

Personally I don't find the idea of 'test first' to be very practical in the kinds of tests that I write. I think the pragmatic programers used the expression 'spike and stabilize' to describe a pattern of development whereby devs get something working then write some tests to support that new feature. They made an analogy of mountain climbers free climbing for a bit then securing everything with the ropes. In order to do 'test first' I would have to decide things like what to call an input field on a web page so that I could write my failing test then code the page. That is too early to make decisions like that. Also, when people say they write a failing test then the impl. are they really writing a single failing test. Often I find I need several tests to cover all the cases. I bring this up because I'm a big fan of TDD but it seems too easy to poke holes in the whole 'write the failing test first' approach.