This is a very nice problem set for anyone interested in the various mechanisms associated with meta-data analysis - they give you their current algorithms and a nice sample data set when you register. As far as I am concerned, the potential prize is just icing on the cake. Good real world problems with well defined success criteria are always a treat.