[Total_Meltdown]

Your part "b" reminded me of a robotics program I did in High School, FIRST robotics - http://www.usfirst.org/ . It's not an exaggeration to say that my success and passion as a developer comes almost entirely from this program. They give these teams of high school kids a real, unsolved, damn hard robotics problem, and you have six weeks to come up with something. It's far more educational than sitting slack-jawed in a classroom, and fun as hell.

And no, there's no cushion for failure. One year I was in this program we had a robot with a drive system that was completely non-functional (we used a worm gear on the drive motors, so the robot basically couldn't coast at all), and as the programmer, it frustrated me almost to the point of tears to not be able to do anything about it. I tried to write code that would simulate coasting by spooling down the wheels slowly, but I wasn't that good at writing code yet, and it didn't work at all. In the end we lost horribly, and our consolation was to do better the next year, the end.

This is the kind of learning I would like to see more of: Work on real problems with no clear solutions, and learn old problems with clear solutions incidentally as the need arises.