Intro to Empirical Software Engineering: What We Know We Don't Know

I thought this was a very thought provoking talk, even if we talk nothing the speaker says for granted.

A year or so ago I had a semi-breakdown over the fact that all the best practices I stood by was more or less just a regurgitation of what some software "guru" had to say on the subject. I didn't know if it was true, but I damn sure believed it.

I believed so hard that my opinions and feelings was almost (if not entirely) religious in nature. Sure, I had some experiences with things that absolutely did not work, some techniques that worked under the specific circumstances I applied them in.

But in general? After 10 years of software development, I can't tell you why I solve a specific kind of coding problem the way I do. I can give you a lot of reasons, but only based on vague hand-wavey stuff such as feelings and experience.

Don't get me wrong; experience is really great. But it's not the same as truth. What works for me might not work for you, and vice versa. And what I believe works for me might in fact slow me down.

And there's no real way for me to know for sure.