Fixing cycle routing anomaly with tricky-to-represent botched TfL infra

Communities United Kingdom
1

I noticed that cycle.travel offered a diversion off the A40 “cycleway” when asked to route eastbound as per this route segment:
https://cycle.travel/map/journey/1038080
I had a look at what this looks like on the ground and how it’s represented in OSM and can see why it’s routing like this. I’m not sure how to tag it so that OSM reflects the half-arsed reality though.
The issue is that the original westbound cycle lane moved from shared pavement to an on-carriageway painted lane along the 40mph slip road. It later routes riders back onto the shared pavement before a busy junction.
Since that lane was created they’ve implemented two way cycling on the pavement with little blue shared pavement signs (a change always guaranteed to please everyone).
The painted lane on the carriageway is represented by the existing tags

cycleway:left=lane

Obviously it’s not suitable for cycling against the direction of traffic.
However to indicate in OSM that cycling eastbound is possible on the two-way shared pavement I’d need to add a separate way with the slip road tags changed to:

cycleway:left=separate

wouldn’t I?
Can I use:

cycleway:left=separate;lane
cycleway:left:lane:oneway=yes
cycleway:left:separate:oneway=no

Or do I just change the tags to make the cycleway separate, create a separate way for it, and pretend the painted on-carriageway lane doesn’t exist any more?
Here’s the OSM way as it is at the moment, recognising the painted one-way lane and ignoring the shared pavement:

There are similar issues on the A4 cycle lanes which are also supposedly two way, but where the transport engineers forgot to take account of the practicalities of riding in one of the two directions. The two way cycle lane routes through bus stops, requiring cyclists travelling contraflow to get slightly too close to riding into oncoming 40mph traffic at each bus stop. Except here TfL forgot to designate the pavement shared space. I think TfL are making it up as they go along.

Would it be worth encouraging the transport engineers to try to represent their designs in OSM so that they notice all the errors they keep making?

2 Likes
2

No responses - I guess it was half just a whinge, but anyway:

  • I was tempted to just delete the cycleway:left=lane and create a separate way for the on-pavement shared space, but taginfo shows that there is a precedent for “lane;separate” (12 instances) so I did that.
  • The trunk link still has sidewalk=left unchanged.
  • I tagged the new way as “highway=cycleway” but arguably it’s highway=footway, bicycle=designated, segregated=no.

Is there an accepted convention for shared space? I believe highway=cycleway implies foot=yes in the UK, but it also feels to me like it prioritises cycling, whereas on shared pavements cyclists and pedestrians usually have equal priority or pedestrians have priority?

Here’s the changeset in case anyone wants to doublecheck/feed back/fix:

3

I think this is a case of “Londonitis”, a particular ailment that afflicts London cycle infrastructure in OSM where the mapper appears to treat it as a schematic diagram rather than, y’know, a map. So you get preposterous situations like this

image
image937×577 163 KB

where it looks like the cyclist has to take a succession of 45° turns to follow the route. cycle.travel has a lot of code to cope with this, but there’s only so much it can do.

In situations like this, I would suggest optimising for the path that the cyclist actually rides (i.e. no 45° turns); using separation= to indicate the lack of segregation; and choosing a highway tag from footway, service or cycleway (with access tags) that best reflects construction and usage.

1 Like