On Mon, Feb 16, 2015 at 11:58 AM, Éric Gillet <[email protected]> wrote:
> 2015-02-16 13:03 GMT+01:00 fly <[email protected]>: > >> There are still cases where forward/backward are useful with P2-routes. >> E.g. a route with a loop and some members used twice but different >> directions. >> > > Shouldn't one just duplicate the stop_positions in the relation, and add > ways twice in order, without roles ? > > Also, as Jo said, the public transport v2 scheme is cleaner as each > variant of the route have its own relation, making loops in routes a rare > case. > You say that, but take a look at routes 108, 222CC, 222CW, 306, and TDT in http://www.openstreetmap.org/#map=12/36.1501/-95.9415&layers=T . This is a problem, a lot of transit systems have equally bizarre routes, particularly when systems are underdeveloped and overstretched like this one. Would be nice if we had a tagging scheme that didn't think of routes as purely linear, or tools that could handle linear objects rationally.
_______________________________________________ Tagging mailing list [email protected] https://lists.openstreetmap.org/listinfo/tagging
