We've been toying with some sort of incentive structure a la SN, but some of us have tried this in the past with the OSM data set and it didn't really work. Maybe this time it's different!