Sidewalk on cyclepath

Hey folks,

while enhancing the pedestrian tagging in my area, I stumbled across “sidewalks” which are also signposted as “cycle path” and part of the local cycle network. Example: Way: 1169028526 | OpenStreetMap

They are usually tagged as highway=cycleway (which I agree on) but also serves as sidewalk, some are also tagged as highway=footway or highway=path. With the latter two, adding footway=sidewalk or path=sidewalk feels ok, though cycleway=sidewalk somehow sounds odd to me and I seems to be not the only one, as those ways have a cycleway=crossing where they cross other roads.
So just wanted to check with you folks, how you dealt with it and have any suggestions? Just use cycleway=sidewalk, even it sounds odd?

1 Like

Explore around the area to see what a typical sidewalk looks like in the neighborhood. If it’s built like a sidewalk, I’d use highway=footway footway=sidewalk. On the other hand, if it’s wider with continuous asphalt pavement and its own signs, then it’s a sidepath. In that case, I’d use highway=cycleway cycleway=sidepath. Either way, I’d use lcn=yes bicycle=designated to communicate the fact that it’s part of a local bike route.

2 Likes

Would concrete work as well? That’s pretty common around here

1 Like

Does this feature bicycle signals, signs and centerlines? Then it’s a cycleway (and may have its own sidewalk). If it’s an extra wide situation with cycleway signage but no centerlines (eg, clearly intended for bicycles but not specific to bicycles such as a multiuse path)? I’d call that a path. If it’s a sidewalk that they’re pretending is a cycleway, I’d call it a sidewalk, bicycle=designated.

Do you have a photograph of what we’re working with?

No, in my area they are just wider than normal sidewalks, no lines, only the signs that those ways are part of the cycle network. They are either asphalt or concrete plates. More or less they were previously a normal sidewalk and got’s widened in a way, that a cyclist can overtake a pedestrian.
Mapillary has an older picture, the small path on the right is gone by now and the wider path is the only one left.

From your links, I would rather think it classifies as sidepath than sidewalk. Though I would not think the surface should matter to distinguish between them.

For sure, that’s out of question.

There is another topic discussing this over the past few years: Documenting in use tag: cycleway=sidewalk - review is welcomed.

I favor highway=cycleway + cycleway=sidepath + foot=designated/yes for these. I agree that cycleway=sidewalk feels a bit weird.

In the other topic I mentioned also adding footway=sidewalk to distinguish shared use sidepaths that function as a sidewalk from bicycle only sidepaths. Perhaps this is redundant though and any cycleway=sidepath + foot=designated/yes can be interpreted as a being also a sidewalk.

Depends. What I meant is that it isn’t built like a normal sidewalk in the area. The surface is often one of the distinguishing characteristics of a sidepath compared to the sidewalks around it. The best sidepaths have all the characteristics of a bike trail, signage and everything, but it can still be a sidepath without all the amenities, if it’s sufficiently different than a standard sidewalk.

Unfortunately, in this case, the street is a one-way frontage road that lacks a sidewalk on the other side to compare it to. But if you go to the beginning of the way, it corresponds to where the surface changed from concrete tiles to continuous asphalt and becomes twice as wide. It’s even more obvious in this case because the sidepath still was under construction back when TomTom rolled through town:

I’d call that highway=path. Don’t forget the name=* to match the street.

That’s what route relations are for.

Sorry, that was imprecise wording on my part. I was referring to one of the green “Bike Route” signs that doesn’t go anywhere in particular. If it has a name or number or a well-defined shape beyond the physical path it’s on, then yes, a route relation would be better than lcn=yes on the way.

That’s the sign that the overwhelming majority of local bike routes are signed with in the US, regardless of length.

Understood. Without more information, it isn’t possible to know whether a relation is appropriate or not. lcn=yes is a fallback when we wouldn’t be able to form a proper route because it’s just being used as essentially a bicycle=designated sign.

(This is really a tiny footnote in the topic. Not sure there’s actually any misunderstanding here.)

I would tag it as highway=cycleway without name.

2 Likes

Thanks folks, I think I will go with highway=cycleway and cycleway=sidepath, not talking about name, access, lcn,…

In general it seems both sidepath and sidewalk are in use for cycleway, just in case someone stumbles across a way looking more like a sidewalk than a sidepath.

1 Like

Some people don’t think that being able to get street names in turn-by-turn navigation shouldn’t be limited only to motorists. Typically names apply to all carriageways of a roadway, of which sidepaths, cycleways and sidewalks are.

I suggest street:name=* if you would like to add an associated street name to a sidepath :slight_smile:

3 Likes

We’ll get around to that around the time that’s routinely done on multiple carriageways. Until then, that’s an inappropriate suggestion.