So I wrote a blog post about anecdote-driven development and I confess that not everyone was swayed by my opinion. Seems that lots of people like to introduce themselves as "hi! I'm captain of the USS Make Shit Up" and then start tossing out "facts" (I can be one of these people when I'm not careful). Here's an antidote:
What we actually know about software development, and why we believe it's true, by Professor Greg Wilson of the University of Toronto.
One bit I appreciated, from another presentation of his, is the debunking of the "best programmers are X times more productive than the worst" myth. He cites "28 times" as a commonly used figure, but I usually hear "10". However, he lists the studies for this information and they usually find the best programmers are only 5 times more productive than the worst and this is consistent with other fields.
I love the idea that we can use evidence rather than guesswork, but I doubt this idea will prove popular any time soon.
It just shows that testing is hard, regardless of context.
Good science is test driven just as good software is test driven (or at least nearby).
