Understandable, and I agree that they definitely tackled the meat of the problem. I think there could be some interesting problems in the computer vision aspect of it, though. Seeing and interpreting the board is subtly different than a straight OCR problem.

Maybe I'm overstating the difference, though.