I came across a few tagging quirks to do with bus access and specifically "service=bus" First of all there's some roads with service=Bus (uppercase B). So that's definitely a mistake.Only exists in a cluster in Manchester ...plus a small cluster in Sao Paulo!:https://overpass-turbo.eu/?w=%22service%22%3D%22Bus%22+global&R
The idea of the service=bus tag is to allow a bus routing system to know that a service road is for buses, with the implication that all other service roads should be not for buses. That was a surprise to me. When I configured a bus routing system (doing this as part of my day job at TransportAPI), it didn't occur to me to disallow buses driving down service roads, but maybe I should rethink that. In any case it seems to me that the access tagging should be regarded as orthogonal to this. What I mean is, while we might tag a road as service=bus, we should also make sure that all these special roads for buses ...allow buses! If they're access=no, then we need to also make sure they are bus=yes (or psv=yes). Sounds reasonable? So I came up with this Overpass query to try to find those problems http://overpass-turbo.eu/s/pDR Not too many. Should we go ahead and fix those things? Quite a small job. I can do it myself actually. There are many other problems (bigger problems!) we could fix with bus data in the UK. For example an extension of this idea would be to check for access tagging on all ways which are part of bus route relations. And of course there's coverage of the bus route relations themselves. That's something we were discussing here as a possible UK quarterly mapping project: https://www.loomio.org/d/uBc7x1Ok/quarterly-project-july-sept2017 Harry
_______________________________________________ Talk-GB mailing list Talk-GB@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-gb
