[Astronouth7303]

One thing you can do is to try to reuse as much code as possible. Over the years, teams tend to develop mechanisms that they reuse to some extent. Do the same with your code.

Like for autonomous. I pulled architecture I wrote in pre-season 2006 but didn't use then. After I got that integrated and working, I was able to write new dead-recon based autonomous modes rapidly.

How much time did I have? Not much on a real robot. But because of the simplicity and my ability to code and debug off a robot, when I got it loaded, it worked like a charm (at least as much as can be expected).

Too simple for me, though. (Give me a gyro! Give me encoders! Give me something nicer than 3 pots and a limit switch!)