Pieren wrote: >was placed on the intersection node itself. >routine engine where routes with traffic signals >are penalized.
I won't be saying anything about the discussed alternatives at this time, but just wish to point out that "this intersection is controlled by signals" when used only on the intersection nodes, can't be straight away used as a constant time penalty for each such node. For example, where two dual carriage way roads intersect, (right hand drive) - a route turning right passes one node tagged with highway=traffic_signals - a route going straight goes through two such interesction nodes - a route turning left goes through three such nodes For some even more complex intersections where carriageways are further divided before the signals, the "wrong count" can be and would be even higher. Even where a single carriage way road crosses a dual carriage, the other road's traffic going straight passes through two such nodes - the other road's traffic only through one such node. I'll let everyone to consider by themselves how the "passed node count" changes on different routes through the intersection, if mappers happen to move the tags away from the intersection node. This shortcoming exists for all intersections where not all roads are single carriageway roads, in different ways. It always need more data, or algorithmically derived guesses. Only when the intersecting roads are both single carriage twoway roads, tagging the signals just before the intersection doubles the "constant penalty" effect, if a router uses the nodes blindly. IMO, therefore, this "time penalties go all wrong" can't be used as a reason why it's always more correct to tag the intersection node. I'd be happy to see someone explain here how their router does something more complex with the highway=traffic_signals nodes. -- alv _______________________________________________ Tagging mailing list [email protected] http://lists.openstreetmap.org/listinfo/tagging
