Pamela Fox of the Google Maps API team recently released a driving directions game based on the Google Maps Driving Directions API. It’s neat, but, especially in areas with which I’m intimately familiar, the routes I choose are far better (more direct, faster) than what Google chooses.
It would be pretty awesome if they built a community out of the driving directions API and used user-generated routes to inform the decisions that the routing engine makes. Essentially I’d like to be an authority on travel in the areas that I am most familiar with (San Diego, Northridge, Simi Valley, Santa Clarita, etc.), including traffic patterns and the “best way” to take depending on various factors.
I suppose one could make that into a third party service and just prefer saved user routes to Google’s designated routes, but the turn by turn directions wouldn’t work and it would probably be a headache in general. Google should take this on to improve their driving directions instead of trusting their (presumably one) fallible data source.
The headache part is probably why it hasn’t been done. Having to wade through endless user submissions that probably wouldn’t even be accurate in the first place and try to translate what they mean with their directions when they miss a street or such. Multiplied exponentially by all the cities in the U.S., even world if they open it up to international directions. I wouldn’t place any more burden on or expect anymore from a company for a service that is already being provided for free, as it is.
Actually I figure the driving directions to a user created route could be extracted from Google’s API by making a bunch of stop-over points.
I also don’t think that Google relies on a source of directions between any two arbitrary points - it would not at all be practical. It is most likely just basic graph theory for the shortest path, possibly with a preference weight for highways. Of course a lot of traffic pattern information is left out, so the resulting directions are not always optimal.