I'm a little disappointed that my proposal to O'Reilly's Where 2.0 was rejected. We would have shown an algorithm to automatically merge vector datasets, premiering a hyperlocal map of Germany.
It's quite simple: primary vector sets from one of the two worldwide leaders is matched with secondary datasets from cadastre, trail maps (or whatever source contains data that isn't available in todays' vehicle maps). 
From what we know, there's noone out there doing the same. Plenty of room to grow, scales like hell, real fun, innovation galore. Would it make a difference to the geospace? Yes, probably.
Don't mistake the business model of prominent conferences.
