tordans
commented
6 months ago The way we think about them in the OSM Parking project is,
- …that
parking=laneis part of the highway, - …but
parking=street_sideis more of a special kind ofparking=surface.
Of course, the distinction is fluid but there are indicators, see https://wiki.openstreetmap.org/wiki/Tag:parking%3Dstreet_side#Distinction_between_street_side_and_lane_parking
The parking=street_side still might have the same surface as the highway which ideally would render it very similar.
Alex rolled out area:highway data for all of Neukölln, which also shows this distinction: https://www.openstreetmap.org/way/1186236308 is the area, the street_side parking is a separate geometry but the parking:left=lane is part of the centerline.
Like https://www.openstreetmap.org/way/888214564. Very unclear how to represent them in osm2streets -- they don't fit into the model of "everything is a road segment with lanes or junction". https://strassenraumkarte.osm-berlin.org/?map=micromap#19/52.47345/13.43859 seems to logically associate them with the street.
Do we snap these to a Road and add synthetic lanes? We'd have to split roads sometimes, or support features that don't occupy the whole width. Doesn't feel like the right approach.
We could just store and render them as areas and maybe provide a logical linking to the road for consumers that want to know "how much parking along a Road". Feels like the cleaner solution. And it feels like it's a gateway towards thinking about pedestrian areas.