I wholeheartedly agree, having also tried both ways. However, too many businesses "spike and move on", and the "stabilize" never happens. This is why people advocate for TDD.
* first, it makes little sense and you follow by rote
* second, you understand and explore
* third you have absorbed and use what is appropriate, naturally
and this is "ri-testing". it's going to confuse shu and annoy ha, hence the defensiveness...
(more generally, i think this is a very useful way of looking at progress. one danger, though, is that it's tempting to think you are ri when you're actually just a ha that's screwing up.) (and apologies for using what i suspect are very carefully defined terms in some marshal art out of context and in grammatically odd ways.)
If that's true then I'm disappointed in HN. This is the best take on testing I've seen; I'm having that experience where you see someone has captured a thought that hadn't quite finished forming in your own head.
Oh, absolutely. It's a fantastic post. I just have a nasty suspicion that a lot of people will read it and think it's an excuse to jump ahead and skip testing too quickly.
