Missing shark teeths for cycleway direction indication

[shadow-identity]

Hey, thank you for the great project!

I just found that in case of

cycleway: opposite_lane
highway: residential
lanes: 1
oneway: yes
oneway:bicycle: no

it renders incorrectly, with cycle lanes on the both sides. While there is only one cycle lane (going opposite). There is no cycle lane in the direction of the line (and it is not described in the tags).

The link to the point on map: https://www.cyclosm.org/#map=19/47.06975/15.43023/cyclosm_live

[Phyks]

Hi,

Thanks for the feedback and the issue! cycleway=opposite_lane is somehow an imprecise tagging, as it does not indicate which lane the cycleway is expected to be.

Most of the case it will be on the expected side of driving (right when riding countraflow for most of the world, left for some countries), but there are many exceptions around the globe. Therefore, we decided to render it on both sides of the road to favour more explicit tagging schemes (lateralized).

We used to have "shark teeths" markers for the cycleways to indicate direction. See screenshot in #260 (comment). That way, it was clear for everyone that cycleways on both side of the street were actually going in the same direction and it was giving an indication that these double cycleways were likely artifacts.

Seems we lost it in latest version, we'll have to look into it.

EDIT: I'm editing issue title to reflect this. Please, do comment if you think having shark teeths to disambiguate is not enough.

[Florimondable]

"shark teeths" has been remoed in an earlier version for optimisation purpose, you were ok with that :)

[Phyks]

True, I remember this now, although I cannot find back the PR about this.

I'm still in favor of removing shark teeths for optimization purposes, but looking at it again I think they might be missing in this specific case of non-lateralized cycleways. Maybe it could be worth restoring them only for this specific case?

[shadow-identity]

Hi Lukas, thank you for your answer. I agree that cycleway=opposite_lane is problematic to determine the exact position of the cycleway if there is no cycleway:lane present.
However, there is another example just near the first one. On the ground this is the same: one-way road with opposite cycleway and without dedicated cycleway in the direction of the main movement of vehicles.
It described as:

cycleway: opposite_lane
cycleway:left: lane
highway: residential
lanes: 1
oneway: yes
oneway:bicycle: no

cycleway: opposite_lane seems redundant for me (since we have cycleway:left: lane set) but still kinda correct. However, it still renders like the road has two lanes.
Maybe in this case we can render the lane from the only one side?

https://www.cyclosm.org/#map=19/47.07034/15.43621/cyclosm_live

[Phyks]

@shadow-identity I don't understand the tagging scheme in your latest example.

If the cycleway and cycleway:left tags describe the same physical infrastructure, they should IMO be merged into a single cycleway:left=opposite_lane (see https://wiki.openstreetmap.org/wiki/Key:cycleway:left), or even better a cycleway:left=lane; cycleway:left:oneway=-1.