Hi guys, I was thinking of this the other day. What if we create a test suite of routing scenarios? The scenarios assume that there's no traffic and toll fees are no problem. We then have a list of pairs of points that we can drive/bike/walk from and to and specify one to two most probable routing scenarios based on our own knowledge. We then check out what the routing softwares (e.g., http://www.yournavigation.org/) decide. By creating such a test suite, we can verify two things:
1. Check that our mapping is correct (access tags, highway classification tags [routing software prefer main roads, for example], oneway tags, etc.) 2. Check that the routing software works as intended (assuming the data is accurate in the first place). For example, I checked out what the best route is going from Quezon Hall in U.P. Diliman to Manila Times, Las Pinas. (This is the route I used to take back in college.) The result from the website above is this: http://www.yournavigation.org/?flat=14.654912&flon=121.064377&tlat=14.44809&tlon=120.984855&v=motorcar&fast=1&layer=mapnik. It's pretty much spot-on except for two things: 1. It made a weird path *clockwise* along Elliptical Road (but it impressively recognized the U-turn slot along Commonwealth Avenue). This should be counterclockwise and probably means that Elliptical Road is missing or has an incorrect oneway tag. 2. It turned left to Taft-Baclaran at EDSA-Pasay Rotonda when it should have gone the Diosdado Macapagal or Roxas Boulevard route. I think the problem is because this stretch of Taft Avenue was tagged as a trunk road instead of a primary road. What do you guys think? We can create such a list at the OSM Wiki. :-) ~ Eugene / seav -- http://vaes9.codedgraphic.com
_______________________________________________ talk-ph mailing list talk-ph@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-ph
