Hi, > > hubert schmid wrote: > > If ways crossing each other without knowing > > anything about the other one, any crossing point has to be > > recalculated. > > OpenStreetMap is topology aware. This means that if two ways cross each > other and the place where they intersect is a routing node, then there > WILL be an OSM node there. If they intersect without an OSM node, then > the intersection is NOT a routing node (e.g. a bridge or tunnel). > > In these cases you can calculate the position where the ways intersect > but this is not of any use for routing.
So my description was not good enough ;) I know that there is always a normal node if ways are crossing but you have always to handle all nodes to find out, which nodes are used by more than one way. It is very hard to navigate in this structure in order to build up a secondary net for routing. In the worst case, you have to store any node of the world im memmory to be sure, that you have found any connection. Introducing route-nodes would help in this case and make the database a little bit more routing friendly. Another idea is to organize the route-nodes in lat-lon-squares, in order to reduce the memmory consumption of the conversion filters. Just an idea for API 0.6 -- GMX startet ShortView.de. Hier findest Du Leute mit Deinen Interessen! Jetzt dabei sein: http://www.shortview.de/[EMAIL PROTECTED] _______________________________________________ Routing mailing list Routing@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/routing
