[Tagging] Feature Proposal - RFC - footway=link

Martin Koppenhoefer dieterdreist at gmail.com
Wed Nov 20 12:49:53 UTC 2019

Am Mi., 20. Nov. 2019 um 11:35 Uhr schrieb Markus <selfishseahorse at gmail.com

> On Tue, 19 Nov 2019, 04:43 Joseph Eisenberg, <joseph.eisenberg at gmail.com>
> wrote:
> > Doing this for every intersection between a path and road, or lower
> > classification road with a high classification road, would be a large
> > amount of extra work for mappers, so it should only be done if there
> > is no other way to get this information.
> However, this is a bit different with ending sidewalks or steps that
> run parallel to the road: tagging the connection with the road
> highway=footway + footway=sidewalk or highway=steps would pretend that
> there were an abrupt change of the direction of the sidewalk or steps
> by 90°, which is a bad representation of the actual geometry (example
> [1]).

Besides, continuing steps up to the highway=* way would distort
> the steps, which is especially problematical with short steps and a
> wide road. For example, continuing highway=steps of 4 steps with a
> length of 35 cm each to the centre of the road that is 8.4 m away from
> the last step would let you assume that one step is 2.45 m long
> (example [2]).

the issue with steps being represented too long is not related to the
proposal of adding a specific subtag. I generally map highway=steps only
for the (approximated) actual projection of the steps (first to last riser
of each steps part) and add highway=footway for the landings and for the
connections on the bottom and top. If you want detail, a step_count is also
useful (add it to the single parts for a good representation). This gives
generally a better representation as compared for example this one next to
your first example: https://www.openstreetmap.org/way/86205958
The method still somehow fails for standard rendering when the steps are
particularly flat, e.g. 1 meter horizontal for every riser but regular step
frequency (I add flat_steps=yes, but it is not very diffuse). Another issue
for 3D stems from situations, where there is both, steps which are also
inclined (but may not be so many), so step_count represents them not so
well (seems less steep than it is).

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20191120/0875a801/attachment.html>

More information about the Tagging mailing list