Hide sidewalks and crossings at lower zoom levels

[planemad]

Micromapped areas have sidewalks mapped as separate ways parallel to a road. This increases the visual noise of the map at lower zoom levels.

_Dashed sidewalks parallel to roads adds unnecessary detail on [this map](http://www.openstreetmap.org/#map=16/50.2549/19.0151)_

Sidewalks are an associate feature of a road and not that important by itself. Ideally we would want the sidewalks and crossings to be less prominent on the map compared to dedicated footways which are more important at lower zoom levels.

A simple improvement would be to display crossings and sidewalks one or two zoom levels later, or make it less prominent by toning down the color.

Taginfo - footway: 400,000 objects (57% sidewalk, 35% crossing)

[dieterdreist]

2015-12-07 10:31 GMT+01:00 Arun Ganesh notifications@github.com:

A simple improvement would be to display crossings and sidewalks one or
two zoom levels later, or make it less prominent by toning down the color.

Taginfo - footway: http://taginfo.openstreetmap.org/keys/footway
400,000 objects (57% sidewalk, 35% crossing)

this is currently not possible, because the key "footway" is not present in
the rendering database. Could be done with hstore available or if a footway
column is added.

[planemad]

@dieterdreist ah, ok. Curious to know how the cost/benefit value is evaluated for adding new keys into the db. Is there a discussion somewhere I could look into?

[polarbearing]

Micromapped areas have sidewalks mapped as separate ways parallel to a road

There is a lot of discussion on the tagging and on local mailing lists, leading to the majority opinion that sidewalks should not be mapped separately unless they are separated by a barrier unroutable for pedestrians, such as grass or fence.

In many cases in the cited map it appears that the sidewalks are just separated by the kerb, where pedestrians can cross the road at any place, thus they should be tagged to the road and not drawn as separate highway.

new keys into the db. Is there a discussion

Actually all issues tagged with Milestone 3.x are waiting for this DB update.

[dieterdreist]

sent from a phone

Am 07.12.2015 um 12:23 schrieb polarbearing notifications@github.com:

There is a lot of discussion on the tagging and on local mailing lists, leading to the majority opinion that sidewalks should not be mapped separately unless they are separated by a barrier unroutable for pedestrians, such as grass or fence.

there is some discussion but there is no such "majority " opinion, at least I ve got a different impression (and I do not advocate drawing sidewalks as independent footways). I believe doing it is mostly tolerated, but unless you do it very complete and detailed, it will often produce worse routing results than without.

[sanderd17]

@polarbearing wheelchair users are also considered pedestrians, while they can't cross kerbs easily.

While I don't particularly like tagging sidewalks as separate way (and certainly not with a value that can confuse existing routers and renderers), no-one can ignore that it's being used (see taginfo in the top post), which is normally the main criterium for inclusion in the style.

[planemad]

If anyone is curious, experimented with rendering sidewalks and crossings for QA purposes.

Play with the Pedestrian QA style

[maraf24]

Tagging footways as seperate way was discussed, voted and approved in 2011
http://wiki.openstreetmap.org/wiki/Tag:footway%3Dsidewalk

Recently Mapbox published tool for checking sidewalk mapping validity:
https://www.mapbox.com/blog/mapping-sidewalks/

[d1g]

There is a lot of discussion on the tagging and on local mailing lists, leading to the majority opinion that sidewalks should not be mapped separately unless they are separated by a barrier unroutable for pedestrians, such as grass or fence.

Such agreement make sense for European cities built in 0-15 centuries (with narrow streets)

https://commons.wikimedia.org/wiki/File:Medieval_street_2_(17129167062).jpg
https://commons.wikimedia.org/wiki/File:Ravenna_025.JPG
https://commons.wikimedia.org/wiki/File:Ravenna_Stadtturm_0609.JPG
https://commons.wikimedia.org/wiki/Category:Streets_in_Italy_by_city

@polarbearing, overall, most footways/separate were welcome during discussions back in 2012. In fact, many of us support and value footways as separate ways. Exact tagging is slightly more complex. If a pedestrian can walk here, then it is hw=footway or hw=pedestrian (with additional tags).

All real barriers should be entered as separate objects.

[jeisenbe]

It would be nice to see some test renderings with sidewalks and crossings rendered later. This might make it possible to render rural paths and footways sooner again.

[JLZIMMERMANN]

Hello is there any rendering to help improving OSM quality about footways ?

[imagico]

Could you be more specific with your question? Is this about general purpose map styles specially treating highway=footway + footway=sidewalk in some form (the topic of this issue) or is is about QA visualization that highlight this kind of tagging. Or is it about any kind of footways (like highway=path, highway=pedestrian, highway=steps nut also implicit sidewalk=* - see #2568)?

[JLZIMMERMANN]

Here you have an example https://www.openstreetbrowser.org/#map=20/44.13702/4.80986&categories=footways,kerbs

another one here https://felt.com/map/Creteil-mobilite-douce-1If72NehSUeTmEYvvw9BZ6A

[imagico]

I am still not sure what your question is about - the two examples you linked to show QA tools visualizing footway features - in this way you answered your own question. If you want to know if we can add something similar in OSM-Carto that would be part of #4723. If this is about interpreting additional taggings w.r.t. footway infrastructure to provide more feedback on this kind of infrastructure in general without explicitly highlighting quality issues - we have this issue (about interpreting footway=sidewalk), #2568 (about interpreting sidewalk=*), #214/#3304 (about more differentiated access restrictions), #1943 (about pedestrian crossings), we also could think about reopening #1290 regarding providing feedback on width=* tagging.

[pnorman]

Returning to the topic of this issue, I've experimented with hiding footways based on footway=sidewalk and it doesn't always work well. When only sidewalks and roads are mapped its fine, but once buildings start to get connected to the sidewalks, it becomes an issue. Those footways should be hidden as well, or else you see disconnected footways

Using overpass turbo, the blue ways have footway=sidewalk, the red ones don't.