The computational geowiki: What, why, and how

Google Maps and its spin-offs are highly successful, but they have a major limitation: users see only pictures of geographic data. These data are inaccessible except by limited vendor-defined APIs, and associated user data are weakly linked to them. But some applications require access, specifically geowikis and computational geowikis. We present the design and implementation of a computational geowiki. We also show empirically that both geowiki and computational geowiki features are necessary for a representative domain, bicycling, because (a) cyclists have useful knowledge unavailable except from cyclists and (b) cyclist-oriented automatic route-finding is enhanced by user input. Finally, we derive design implications: for example, user contributions presented within a route description are useful, and wikis should support contribution of opinion as well as fact.