Talk:Tag:cycleway=track

Protected bike lanes as separate ways

This article simultaneously says that a protected bike lane can be mapped as a separate way tagged highway=cycleway and that the combination highway=cycleway cycleway=track is invalid. Should such ways lack a cycleway=* tag? If so, how to distinguish these protected bike lane ways from dedicated bike path ways (e.g., in the middle of a park)?

The guidance about highway=cycleway cycleway=track seems to be focused an old practice of automatically tagging every dedicated bike path with cycleway=track, but this is actually a valid tag combination when mapping protected bike lanes as separate ways, right?

 – Minh Nguyễn ^💬 18:29, 8 February 2022 (UTC)

cycleway=* is a property of a road - whether/what kind of cycleway is present for this road-way. It is a (a bit) ambiguous short-hand for cycleway:left=* + cycleway:right=* set to the same value. highway=cycleway defines a way as a cycleway. Adding cycleway=track to a highway=cycleway thus is like saying that this cycle track has a cycle track (left and right). --Westnordost (talk) 18:34, 8 February 2022 (UTC)

@Westnordost: Yet cycleway=crossing also exists. In this sense, cycleway=track is just iterative refinement of highway=cycleway. I can understand how this overloaded usage of cycleway=* could be confusing to some data consumers, but on the other hand, there should be a tag to distinguish a protected bike lane from a dedicated bike path. oneway=yes sometimes works for that purpose, but there are two-way protected bike paths within the roadway, and there are also one-way dedicated bike paths away from any roadway. Incidentally, one-way cycle tracks are being mapped, but not in such high numbers that an alternative couldn't be introduced. – Minh Nguyễn ^💬 18:42, 8 February 2022 (UTC)

Note that it was primarily added as result of Potlatch 1 preset bug, see https://help.openstreetmap.org/questions/65366/highwaycycleway-cyclewaytrack I am not aware about people deliberately adding it. Though maybe someone is using it in way you describe? Mateusz Konieczny (talk) 09:16, 9 February 2022 (UTC)

@Mateusz Konieczny: Yes, I'm aware of that Potlatch bug, but this query generally finds deliberate usage of this tag on protected bike lanes, which are generally one-way. – Minh Nguyễn ^💬 09:50, 11 February 2022 (UTC)

There's cycleway=sidepath together with path=sidepath. --- Kovposch (talk) 07:51, 9 February 2022 (UTC)

@Kovposch: Thanks, it's inconsistent with cycleway=track on roadways but adequate to make the distinction I was asking about. – Minh Nguyễn ^💬 09:50, 11 February 2022 (UTC)

I use the more flexible is_sidepath=yes plus subkeys. I do not like the idea of adding cycleway=* to highway=cycleway, especially, if the same values are used as additional tag for the main road. --Skyper (talk) 15:46, 14 February 2022 (UTC)

To avoid the described issue with highway=cycleway + cycleway=track, I use cycleway:type=* in my area (still undocumented and rarely in use). Typical values are track, crossing, sometimes also lane or link (for "virtual" connections).

To specify protected bike lanes, you can have a look at the Separation Proposal. --Supaplex030 (talk) 08:43, 13 February 2022 (UTC)

What's the difference with cycleway=*? Seems arbitrary. There's no key conflict here. Although footway=sidewalk is not used on roads, there have been ideas eg footway=lane before. --- Kovposch (talk) 12:23, 13 February 2022 (UTC)

Whether there is a difference or not depends on the interpretation. As described above, cycleway=* is traditionally defined as the existence of a cycle path, not as a type. (Since footway=* is not in fact used in this sense, this problem does not arise there). We have discussed this for a while in our local community and decided to prefer cycleway:type=* for now. When it is clear that highway=cycleway + cycleway=* is interpreted correctly [Edit: Or better: not wrong, i.e. not as a cycle path with a cycle path - this is what the JOSM template does, for example], the tagging can easily be moved again. Personally, I would also prefer just cycleway=*. --Supaplex030 (talk) 12:39, 13 February 2022 (UTC)

Well, footway=* was used for both in the past but replaced by sidewalk=*. I prefer the same for cycleway=* as it adds less honeypots for misinterpretations. --Skyper (talk) 15:46, 14 February 2022 (UTC)

Far too subjective

The given criteria of "separated from the road by curbs, parking lots, grass verges, trees or another physical barrier" makes it a separate entity. It requires the *primary* tag highway=cycleway. "but is running parallel to and next to the road." < How much divergence is required to make it not "parallel"? How far way before it's not "next to" the road? Are there any renders which show cycleway=track? Even the main page cycle maps don't display. --DaveF63 (talk) 17:45, 27 April 2022 (UTC)

This tag does not seem to make any sense

I am super confused why this tag was invented and how to use it. How should any software know in which direction the cycle way can be used? Whether it is compulsory or optional to use? How is a map or a routing software supposed to present to the user what the surface and smoothness of the cycleway is and what is that of the road? It seems like an antipattern to me to use this tag rather than mapping the cycleway separately. Candid Dauth (talk) 13:13, 19 July 2026 (UTC)

All detail attributes can be specified using the cycleway=* prefix (e.g., cycleway:left=track + cycleway:left:surface=paving_stones + cycleway:left:width=1 + cycleway:left:traffic_sign=none). This also applies to the direction of travel, which corresponds to the defaults depending on the side and direction of traffic (e.g., cycleway:right=* indicates travel in the direction of the line in countries with right-hand traffic). Nevertheless, separate geometry has many advantages and often replaces the “simplistic” centerline tagging sooner or later. Supaplex030 (talk) 18:47, 19 July 2026 (UTC)