Day 2 started with a nice, greasy hotel fry-up, setting me up nicely for the day. First up was the keynote from Robert Martin, talking about clean code. See the slides: a very interesting and, most important for me, useful talk. A brief digest: you don't need comments, don't comment out code, and don't write code you're not proud of - do it the right way the first time. Otherwise the code just rots.
Lunch (butter chicken and rice, yum).
"Working with legacy code", wasn't what I was expecting (see the slides). Legacy code, for Michael Feathers' talk, was just old code you wrote before you started writing testable code. You still have the source and you're free to start modifying that code. I didn't like the way he was adding new paths throuugh the code just for testing. His presentation skills also sucked. I had been hoping that it would be about insulating your code from the problems that come up from old code that you have no control over.
Finally, Erik Doernenburg showed us ways to visualise code quality. Executive summary: plot your checkstyle numbers on graphs to make it easier to see dodgy areas, because your brain is better with pictures and patterns than numbers.
After that there was supposed to be developer networking and a chance to meet the speakers. Fuck that, who wants to spend all that time talking to geeks about optimistic locking and functional programming, even if there is free beer? I'm a very anti-social person. So I ran away to the pub on my own (see "beer" post below). Awesome.