Oops, I must take this back. Further research indicates it's not a data error on the intersection. This becomes evident by choosing the shortest route instead of the fastest... Probably Gosmore thinks that the speed advantage of the trunk highway will arrive you quicker at the destination then using the shorter primary route. This most likely indicates that the weightings in the elemstyles.xml file for Gosmore should be refined.
Lambertus wrote: > This is where some of the strengths of this webservice implementations > shows. You can place to and from markers on multiple places around the > route to see where the problem originates (You don't need to clear the > route entirely, just place a new to or from marker. > > So I've investigated this and found the most likely problem. Please have > a look at the following two routes: > <http://tile.openstreetmap.nl/~lambertus/routing-world/?flat=51.031219&flon=-1.076628&tlat=51.058047&tlon=-1.265553&v=motorcar&fast=1> > > <http://tile.openstreetmap.nl/~lambertus/routing-world/?flat=51.014545&flon=-0.924744&tlat=51.058047&tlon=-1.265553&v=motorcar&fast=1> > > These two routes seem to me an indication for a data error on or near > the intersection, probably disconnected nodes. I hope this helps. > > Jack Challen wrote: > >> First off, well done! It's very, very impressive. >> >> However, here's a route which seems a bit sub-optimal: >> >> http://tile.openstreetmap.nl/~lambertus/routing-world/?flat=51.014545&flon=-0.924744&tlat=51.058047&tlon=-1.265553&v=motorcar&fast=1 >> >> >> IMHO it should travel almost directly west along the A272 (which appears to >> be connected at both ends). >> >> cheers >> jack >> >> >> >> >> >> > > > > _______________________________________________ > talk mailing list > talk@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk > _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk