Cycleway "lane" or "track" for Protected Bike Lanes?

Yeah I’m not sure if I’m joking or not either so that makes two of us. I’m actually a software engineer at a Fortune 5 company so I’m seeing a lot of things all at once happening in OSM. For instance we have detail creep in that a lot of people want to use OSM as a repo for geodata that goes well beyond streets. At the same time trying to keep OSM super simple also causes us to do truly off things like separate roads in two must because they are divided instead of treating them as one thing. In the name of backwards compatibility we exclude who type of “traffic and lanes” from being “traffic and lanes”.

One of the more interesting things I’ve done recently is use JOSM with the Styles/Lane_and_Road_Attributes – JOSM add-on and suddenly with an overlay I could see how much detail was both missing and in many cases completely wrong because as tagging gets more and more complicated it become easier and easier to screw up even if you have a programmer brain with a propensity for minutia.

So no I don’t think the above is THE solution but it also demonstrates THE problem as well. If we are going to capture the detail we want and/or need we need tools to make the mapping process better. I think the more we actually try to visualize things in OSM it will start to expose its weaknesses. Like why this post started with confusion about if cycleways are tracks or are lanes and how I can use OSM and wiki to argue they are both and/or neither depending on who you ask and both answers break the moment a bike lane falls between two traffic lanes!

1 Like

The physical separation criterion is good to uphold regardless. Even if we completely ignore complexity/maintainability, it’s still a better idea to have various roadways/objects of a street as components in a street relation, than tagging everything on one way. In this approach, you indeed have one object to refer to the street, but you can also refer to individual components with their spatial positioning correct, which is needed for detailed renderers at high zoom levels.

1 Like

You’re not wrong, but here’s the actual current tagging from the location of the photo at top of this topic:

bicycle:lanes=no|no|designated|no|designated
cycleway:lanes=none|none|lane|none|lane
cycleway:right:separation:left:lanes=||vertical_panel||solid_line
cycleway:right:separation:right:lanes=||solid_line||kerb
turn:lanes=left;through|through|through|right|right
vehicle:lanes=yes|yes|no|yes|no
width:lanes=||1.7||1.7

Tagging like this is a nightmare to work on without the assitance of various JOSM styles and plugins dedicated to lane tagging, and even then it’s still pretty confusing.

For me, the protection in “protected bike lanes” is that cyclists are protected from cars. The bollards are primarily not there to stop cyclists to change onto the car lane, but to prevent cars to change onto the bike lane.

I’m late to the party, but it might be good to compare existing examples? These wand separated lanes are increasingly common in the UK. I think the problem they solve isn’t so much the risk of a car veering into the lane (they pop out when struck by a vehicle. They help a bit, like painted lanes, by encouraging drivers to leave space, though some argue painted lanes encourage drivers to drive closer to cyclists because they think they can pass as close as they like as long as they’re in the “car” lane) but they stop cars parking in the bike lanes, saving the cyclist from having to regularly swerve in and out of traffic and judge the door zone. We’ve reached a point in London where many painted bike lanes are just used as car parking because they aren’t enforced.
OSM already recognises in common tagging practice a classification difference of cycleway:lane=advisory and cycleway:lane=mandatory which locally is the difference between a lane demarcated by a dashed line vs solid line respectively. There have been suggestions recently that parking on mandatory lanes will be enforced with fines.
The wanded lanes have been tagged in London as cycleway:lane=exclusive and I think that might be the right distinction?
Some examples:
Way: ‪Uxbridge Road‬ (‪4043506‬) | OpenStreetMap (advisory)
Mapillary (real world example of advisory cycle lane)

Community only allows two links per post for new users, so here’s the OSM and real world example of a wanded bike lane!

Way: ‪High Street‬ (‪1065595700‬) | OpenStreetMap (exclusive)
Mapillary

Quoting the wiki, cycleway:lane=exclusive is used for cycle lanes “strictly reserved exclusively for cyclists, segregated from the car lanes usually through a continuous line. In most countries, this implies an obligation for cyclists to use it.” To explicitly distinguish Protected/Wanded/… Bike Lanes with physical separation, there is the separation proposal.

But obviously the community is extremely undecided whether lane or track is the more appropriate cycleway value for this - currently the vote in the first post is exactly 50-50 :joy:

Okay, this is calling for extreme measures. Write a proposal to deprecate cycleway=track. Physically separated cycling tracks should always be drawn separately. If you want to map the cycleway as part of the road because it is not physically separated and/or part of the road surface, use cycleway=lane.

To be honest, I am not really sure if I am kidding or not.

But the problem will remain - because then we do a 50-50 vote on whether protected bike lanes should be mapped separately or not :wink:

Yes, but edit wars might become less likely and have less of an impact. It is easy to change cycleway=lane to cycleway=track and make width and all :lane tags incorrect by that. Drawing and deleting a separate way is a bigger effort.

And I also think that it would have a psychological component. If it feels correct to draw a separate line, then it might really be separated enough to be its own track.

1 Like

“In most countries, this implies an obligation for cyclists to use it.” - and here we get into country differences. In the UK I’m not aware of any cycleways that legally obligate cyclists to use them over the main carriageway. I find it confusing enough that in the UK, highway=cycleway implies “foot=yes”
I suppose the OSM rule would be to tag what’s there and let routing engines decide how to interpret them, which still doesn’t clarify the distinction between lane/track for infrastructure that feels like it’s between the two.

Only ones I can think of are the ones avoiding the A55 tunnels between Conwy and Bangor.

A cycle lane like in the picture below, which is not only separated by markings but also by e.g. posts or bollards, might therefore be a “track”. I would like to get a sense of the community’s opinion on this, to sharpen the definitions for this case.

This is neither a lane or a track, but should have its own way tagged as:

highway=cycleway
lanes=1
oneway=yes

at a minimum.

In the case the cycleway on the image above shows a “track”, I wonder whether we see a “track” in the middle of the road and a “lane” on the right street side in this image :

For the part right of the bollard row…

highway=*_link  # whatever level the link is attached to
cycleway=lane
lanes=3
access:lanes=no|yes|no
bicycle:lanes=designated|yes|designated
turn:lanes=through|right|right

=track means physical segregation, =lane means paint only. Wands and bollards are physical segregation. I can’t see any other meaningful way of defining the difference between track and lane.

Therefore example 1 is a track because it’s physically segregated from motor traffic, and example 2 is a lane because it isn’t.

By all means put the nuance in additional tags like cycleway:separation=wand.

5 Likes