Our little group of friends (me, Vince, Mark, and Daniel) have a book club and yesterday we discussed Hillel Wayne’s What we know we don’t know, a talk about things that we do know about how to develop software well, and the vast vistas of ignorance about that.

Short summary: the talk content is good, you should watch it or at least read its web page. It’s a short, dense talk, and I’m not going to summarize its main points. It’s also likely to be an interesting talk to anyone who develops software professionally.

Some points that came up in our discussion:

  • there are actual, scientific, peer-reviewed papers on results of software engineering research now, and that is an excellent thing
  • individual papers are not enough, in any branch of science, and it’s good to see results are getting reproduced and challenged
  • the talk emphasizes that testing is good - which we were glad to hear, even though the talk discusses mixed results about whether specific approaches to testing are better than others
  • we enjoyed learning about results that code smells and bugs correlate, and also that fixing code smells doesn’t fix bugs
  • our experience is also that code reviews work, but the challenge is getting people to do reviews and to present changes in a way to make reviews easier to do