Proposal talk:Pedestrian crossing as an area

From OpenStreetMap Wiki
Jump to navigation Jump to search

Redefinition

This proposal currently discounts area:highway=footway footway=crossing on the basis that it would redefine area:highway=footway. But how is that any different than highway=footway footway=crossing redefining highway=footway? Tag:area:highway=footway#Example tagging of connected objects does recommend against area:highway=footway footway=crossing, but only because it favors adding junction=yes on the overall intersection area. In other words, this proposal inherently redefines that page anyways.

It seems to me that area:highway=footway footway=crossing would be the most appropriate counterpart to a crossing mapped as a highway=footway footway=crossing way, while area:highway=crossing would be a fallback for when a crossing can only be mapped as a node but not a way. Can the latter situation ever occur?

 – Minh Nguyễn 💬 06:28, 24 January 2022 (UTC)

I fully agree, and indeed made precisely the same point in the forum. --Waldyrious (talk) 13:30, 26 July 2024 (UTC)

Cycleway–footway crossings

How would the proposal distinguish the following marked crossings:

  • Footway across a cycleway [1]
  • Footway across a ped-bike path [2]
  • Cycleway across a footway (can this happen?)

Do these distinctions matter? In other words, do the access tags only indicate what modes of transportation may occupy the crossing, or do they also serve a role similar to footway=* and cycleway=* in ensuring continuity with the adjoining paths?

 – Minh Nguyễn 💬 06:47, 24 January 2022 (UTC)

Multiple areas per crossing

The proposal states: "Note that in case of crossing having two separate parts, one for cyclists and one for pedestrians it should be mapped as two area:highway=crossing - each with its own access tags." I believe this means that there can be multiple area:highway=crossing polygons for one highway=crossing node, and that the simple logic of identifying the relationship between polygon and node (by checking which polygon contains the node) won't always work. It would be good to define an unambiguous rule for finding the related node in this case. --Tordanik 12:01, 5 February 2022 (UTC)

What about "merge adjacent areas representing crossings (consider layer=* tags if present), find highway=crossing node within it, in case of multiple ones prefer one with the same layer values"? Mateusz Konieczny (talk) 13:33, 6 February 2022 (UTC)