Most Coding Dojos follow 2 main formats: Kata and Randori. Both formats try to give as much information as possible to the audience by evolving solutions from scratch. This is great when dealing from Novice to Competent level but tends to get slightly less interesting to people at a higher level. The Kake format goes beyond competent by adding an extra challenge. It increases dynamics by putting people into a more “real” world exercise. Participants deal with unknown code bases and have to evolve from them. This clinic will present the format and initiate a real session with the attendees.

Experimentation is one way to gain insight into how processes perform for a team, but teams rarely do experiments, fearing such educational excursions will incur extra costs and cause schedule overruns. When facing a stalemate concerning pair programming, one team performed a lightweight experiment evaluating pair programming and programming alone with inspection. Through the experiment, the team learned that pair programming was faster than programming alone, required less effort, and had more predictable quality. Lightweight experimentation is easy, cost effective, and fun.

How many doctors pair diagnose? How many architects pair draw? How many journalists pair write their articles? If they don’t need to pair work, then why should we? If it works that well for us, then why don’t they do the same thing?

I don’t believe in pair programming. If you think we need it because it’s too scary to let developers sit alone and code, you have a fundamental problem. Why don’t we require each developer to deliver high quality code, even if he works on his own, instead of setting two sloppy developers together and hoping they’ll correct each other?