A while ago I asked about the best way to handle an unsigned segment of a signed route. Clearly the ways themselves would be given unsigned_ref=XYZ, however it's the route relation that gets fuzzy. I don't think the discussion really went anywhere definitively
I don't know if most routers derive route information from relations, however we certainly seem to be moving in that direction. The unsigned_ref tag is perfectly valid in a relation, however that would apply to the *entire* route. We can't just opt for the role "unsigned" either, since that makes it impossible to use forward/backward/north/south/etc. roles. Consider US 19 Truck in Pittsburgh. It starts in Mt. Lebanon, running along Liberty Avenue to PA 51. It then turns onto PA 51 until it meets the I-376/US 19/PA 51 interchange at the south portal of the Fort Pitt Tunnel. Here the signage stops, however a sign directs drivers to follow I-279 to continue on 19 Truck. Once out of the tunnel on the north end, it follows I-279 to the McKnight Road exit, where signage resumes. The route then turns onto McKnight Road, running along that to its end at US 19 in McCandless Township. Presently there are two route relations for this route. One for both signed segments, with a gap in the middle. The other is for the route as a whole, including the unsigned segment in the middle. I suggest changing it up. Have one relation for each of the three segments: the south signed segment, the middle unsigned segment, and the north signed segment. They would then be unified by a super relation, much like the ones that unify the relations for Interstates and US routes in each state. How would this be as a way of handling this?
_______________________________________________ Tagging mailing list [email protected] https://lists.openstreetmap.org/listinfo/tagging
