On Tue, Jan 11, 2011 at 3:26 AM, Steve Bennett <stevag...@gmail.com> wrote: > On 11/01/2011 12:20 AM, Richard Mann wrote: >> >> The user who'd prefer to use highway=cycleway ways doesn't >> know that the cycleway=track is a duplicate, but routers only have to >> give a slight preference for highway=cycleway over cycleway=track to >> use the "right" one (and even if they use the "wrong" one, it doesn't >> much matter anyway). > > IMHO, this is being extremely optimistic about the powers of a router. So > you're saying that a router could see the "adjacent=yes", then locate a > nearby cycleway which is "adjacent" in some way, and conclude that the two > refer to the same thing? > > My suggestion: let's get a relation happening, asap. Something like:
No, I'm saying the router doesn't need to know. If they have two almost identical paths in their network alongside one another, it doesn't matter which they pick. As for a relation - it's too complicated. What happens if the ways are different lengths. What happens when someone chops one way in two because of something else happening (etc etc etc). Relations are inherently complex: only use them for things that people will readily understand. Richard _______________________________________________ Tagging mailing list Tagging@openstreetmap.org http://lists.openstreetmap.org/listinfo/tagging