[mathking]

First I would just like to stress that in my view, the posters who talked about deciding on strategy first are right on target. On the question of how much freedom to give the students a lot of it depends on your team culture. Our team gives the students a lot of leeway. Probably more than most. That's OK for us because it is part of our team's normal operating procedure. We have more than once let students wander down a path I suspected would have to be changed, because all things being equal, they learn more when they fail. (To be fair, I try to always have an implementable alternative ready to go in these cases.) But it is also true that quite a few times something I and the other mentors thought would not work actually did work very well. There are some very smart, creative kids doing FIRST. But you have to be wary of this approach too, if the kids and/or mentors aren't used to it, you can fall behind rapidly or have a lot of hurt feelings and general disgruntlement.

What helps us is having student leaders who help guide the discussion, design and prototyping process. In particular the student leaders work on the proof of concept stage. Once that is passed and we settle on a general design the mentors take a more active role. At least most years. Since we have mostly college student mentors, our mentor knowledge base varies. For example last year we had really only one mechanical engineering mentor, and she was new to ME. Fortunately our student engineering director was one of the best FIRST students I have ever had, and he basically filled the role of a mentor for mechanical stuff, allowing me to keep to the big picture view. In any event, if the student leaders are good, they will help distill all of the ideas into some good, solid designs the team can discuss and choose from.

Which brings me to the big picture view. When you give the kids more freedom, you have to make sure that someone among the mentors is both keeping track of the big picture, and making sure that the kids are thinking about how everything is going to work/fit together on the robot and strategy wise. It is really easy for the designers of each component to be drawn down the path of "it's just one little change" so many times that things no longer work together. (Heck, this happens with engineers too, just less frequently.)

Just remember that you are not looking for the right way to design a robot but a way that is right for your team.