New quest: Tag sidepaths with `is_sidepath=*`

[mcliquid]

General

Affected tag(s) to be modified/added: is_sidepath=*
Question asked: Does this path run along a road or is it an independent path?

This quest was proposed upstream in StreetComplete (streetcomplete#6722) and closed as “not planned”, mainly due to concerns about survey value and potential spam. As SCEE allows more experimental and niche quests, this proposal is brought forward here for consideration.

Checklist

Checklist for quest suggestions (see guidelines):

• 🚧 To be added tag is established and has a useful purpose
  is_sidepath=* has been proposed since 2015 (RFC 2022) and is in active use (~50k ways). It is used by applications such as cycle.travel, Radverkehrsatlas and Cycling Quality Index.

• 🤔 Any answer the user can give should have an equivalent tagging
  Answers map directly to is_sidepath=yes or is_sidepath=no.
  For highway=footway an additional option “This is a sidewalk” can set footway=sidewalk instead.

• 🐿️ A survey is necessary
  Whether a path is perceived as running along a road (with clearly perceptible traffic context) or as an independent route usually requires on-the-ground observation and cannot always be reliably determined from aerial imagery.

• 💤 Not an overwhelming percentage of quests have the same answer (No spam)
  Both roadside and independent paths are common. Spam can be reduced by restricting the quest to paths near and roughly parallel to motor-traffic roads.

• 🕓 Applies to a reasonable number of map data (Worth the effort)
  Separately mapped cycleway, footway and path features are widespread globally.

Ideas for implementation

Element selection:

Ways without existing is_sidepath=* and matching one of:

• highway=cycleway
• highway=footway and footway!=sidewalk
• highway=path with bicycle=* and/or foot=* and path!=sidewalk

Exclude:

• Ways already tagged with is_sidepath=*
• Ways clearly intended for motorized traffic
• area=yes
• highway=* ending in _link
• access=no|private
• bicycle=no|private

To address upstream concerns about mass-tagging of remote paths, additionally restrict to paths that:

• are within a defined distance (e.g. 20–50 m) of a motor-traffic road, and
• are roughly parallel to that road (to avoid crossings or short connectors).

Metadata needed:

No country-specific legal metadata required. The classification is based on physical and perceptual context, not traffic law.

Proposed UI:

Question:
“Does this path run along a road?”

Hint:
“A path runs along a road if you would describe it as going along that road and the road and its traffic are clearly perceptible. If the path provides an independent route (e.g. through a park, forest, or green corridor), select ‘Independent path’.”

Answer options:

• “Runs along a road” → is_sidepath=yes
• “Independent path” → is_sidepath=no
• (for highway=footway) “This is a sidewalk” → footway=sidewalk
• “Not sure” → no change, option to hide/comment with photo

Alternative: Super Simple Yes/No Question

[Helium314]

are within a defined distance (e.g. 20–50 m) of a motor-traffic road

This seems quite far, wouldn't you usually expect it within like 10 m?
Though even with this an overlay might be a more suitable solution.

“Not sure” → no change, option to hide/comment with photo

What's the difference to the usual "uh... -> leave note"?

[mnalis]

are within a defined distance (e.g. 20–50 m) of a motor-traffic road

Those numbers probably come from my completely wild guessing here: streetcomplete#6722 (comment) 😅

This seems quite far, wouldn't you usually expect it within like 10 m?

10m "from the end of the road to the start of the sidewalk" sound reasonable enough to me (even about 5-7m might be good enough for many cases)...

But how far is that from the centerlines (i.e. ways as drawn in OSM)? E.g. if the road has 2+2 lanes, and each lane is say 3.75m, add a protective grass/tree strip, and a width of sidepath too (esp. if it is combined sidewalk+cycleway) -- it grows, and needs to be taken into account.

The comment above links also to streetcomplete#4494 which had some more serious calculations in the code about how to calculate the "distance from edge of sidewalk is to the edge of the road" given that we can only directly see in OSM "distance from the centerline of the separately mapped sidewalk to centerline of the road".

So we can resurrect that code... Or perhaps such complexity is unwanted, and we can go with some hardcoded values (like e.g. 20m between centerlines? Or whatever makes sense given the problem above)

Though even with this an overlay might be a more suitable solution.

Yeah, I too think Overlay would be more suitable... But if Quest can reliably eliminate those unwanted forest/park/playground/etc. footways, it might also be fine. The problem is we can't really test it if would is good enough until the quest is written 🤷, and it would be a shame if it turned out it just can't be made good enough and that effort was wasted and rewrite as Overlay was needed after all 😓

[Kovoschiz]

The is_sidepath:*= family should be considered together as a whole. is_sidepath:of= has been questioned before. The unwieldy is_sidepath:of:name= is losing favor to street:name= now.
“Runs along a road” is over-simplified. It needs to be adjacent, associated, and mostly accessible on the same level when obstacles are removed. https://wiki.openstreetmap.org/wiki/Key:is_sidepath#Edge_Cases:_When_it's_a_sidepath_and_when_it's_not?