15:th of March we had 15 participants, and this time we chose to split into two groups. Since we had an extra laptop (a MacBook by the way), an extra video projector and a large enough room this was quite a simple way to ensure that everyone would be allowed to take active part in both discussions and coding. Training just isn't training if you are just watching others… We ran the Counting Lines Kata in both groups by the way. Our biggest insight from that session was that you can't take too small steps. As long as you force the implementation forward by adding tests that should always be able to run it does not matter if you solve a "whole" feature, or just a part if it. The important thing to remember is not to write tests that checks the implementation, but rather the behaviour of the class.
The 29:th of March we had 12 participants, and this time we went back to a single group. We ran the Counting Lines Kata again (I'm surprised that nobody even suggests that we should change the Kata at the retrospectives at the end of each session. I'll never get to use all my good arguments for why we should be training on "known ground"… soon I'll change Kata just to see if anyone wants the old one back!). This time, and actually the previous time, we spent a couple of minutes discussing design in the beginning. It's feels that the group is more satisfied if we put a couple of design ideas on the table, and then remove all but the simplest that could work, instead of just stating that the first one we think of is the simplest (and will work).
Tomorrow it's time for the next session. See you then?
/ Andreas
