Talk:Cycling in the United Kingdom

From OpenStreetMap Wiki
Jump to navigation Jump to search

Completion / relations vs ncn_ref

Note: these routes appear to be more complete than the below would indicate. For example, NCN route 1 is much more than 15% complete - perhaps closer to 75%. But some of the routes are tagged as "ncn_ref=1" rather than using the relation, and there doesn't appear to be any easy way to convert them. The best way I have found so far is to load up josm, download data for the locality in question, use the "Search" function to find "ncn_ref:1" - this will match any route containing 1 (is it possible to do an exact match?). To exclude other routes that match, eg 14 and 10, do "ncn_ref:1 -ncn_ref:14 -ncn_ref:10". Then add the selection to the relation and delete the ncn_ref=1 tag. Unfortunately this procedure is slow and slightly error prone. (comment added by User:ChrisB

A few separate issues here. Firstly, ncn_ref isn't wrong - it's no longer the "preferred" way of doing things, but is still perfectly valid.
Secondly, I think it's mostly a case of people just not updating the completion tallies - i.e. just a wiki content issue, nothing about the actual map (which is the important bit!). Feel free to update. In particular I suspect we could lose two of the three columns entirely now.
Finally, the easiest way I've found to change an ncn_ref-tagged route to a relation is just to select the cyclemap as the Potlatch background layer (so you can see which bits are NCN routes), then go along bit by bit adding to the relation: presumably you can do something similar in JOSM. --Richard 18:25, 23 November 2008 (UTC)

On Road Routes

A few bits of the NR5 around Oxford (and maybe others, but I know of these ones) are on road. What's going to be the best way to label these? (eg NR5 runs along Kennington Road, do we label it as "Kennington Road / Cycle Route 5", or something else)

I have the same question. I wonder how road numbers are being labelled? I think there should be a separate label, for example: name=High Road, Fosse Way, A43, NR8; roadNumber=B404 ; cycleRoute=NR5; A road can have many names so the name label should probably be free form with no standardised structure or format, while there should be specific labels for standardised data. 80n 13:11, 27 Feb 2006 (UTC)
I've tagged a couple of ways with an 'ncn_ref' key/value. For example, ncn_ref=NR45 means National Route 45. RR40 would be Regional Route 40. --Richard 18:44, 14 May 2006 (UTC)

On Road vs Off Road

Sustrans have different codes for bits of the route, to indicate if they're on road / off road etc. Do we want to do something similar explicitly, or just try to do it implicitly?

As far as I can see Sustrans uses the following colour coding:
  • red - on-road
  • green - traffic free
Are there others that I couldn't find?
The existing class label (class=path or class=primary|secondary) can describe this information. However, I'm unsure about how to describe a cycle route that is on a shared use pavement. It's not strictly on-road, but it's not a separate path on the map. Sustrans shows these as traffic-free (green). Maybe we need a modifier to the bicyle permissions label (bicycle=no, yes, markedLane, sharedFootpath). 80n 14:06, 27 Feb 2006 (UTC)

Sustrans also have light green and orange routes, which are the equivalent of green/red for non official routes. In terms of routes, I'd say we have:

  • on road, no marked cycle lane
  • on road, cycle lane marked along edge
  • on road, barrier dividing cycle lane from road
  • on pavement
  • cycle lane runs parallel to road, but on a different bit of tarmac
  • cycle route on track / dedicated cycle path
  • cycle route off road (eg dirt path)

Gagravarr 25:53, 27 Feb 2006 (UTC)

Proposed Labelling Scheme

Here's a proposal. Suggestions, criticisms, comments and counter-proposals are welcome:

  • class - defines the type of road using the already established scheme (class=path|primary|secondary|motorway etc).
  • bicycle - a rights/permissions label (bicycle=yes|no|markedLane|pedestrianShared). The implied value would be yes for roads (except motorways) and no for paths.
  • bicycleRoute - a label identifying the route number according to the region's classification system (bicycleRoute=NR1|NR2|...|The Surrey Cycleway|Surrey Cycle Guide No 5|...).

Even with something so simple there are problems and issues:

  • A single carriageway road could have a marked cycle lane on one side, and a pedestrian shared pavement on the other side.
  • A path could be part of more than one cycle route (NR1 and The Surrey Cycleway and Cycle Guildford).
  • The presence of a bicycleRoute label implies bicycle=yes, does that make the bicycle label redundant.

80n 14:33, 27 Feb 2006 (UTC)

Is "(class=path; bicycle=yes)" still the correct way to tag these?, as it still says to at the bottom of the page. I've just added the ncn_ref tag to the roads it follows. A road allows bicycles by default and isn't a path. Ben. 15:26 25th January 2007 (UTC)
See the Cycleway key, not sure if it is the correct key to use. The "class" key is depreciated. The "ncn_ref" key appears to be the NCN reference key to use, according to Key:reference. - Bruce89 17:40, 25 January 2007 (UTC)

2020 Abandonment of some routes by Sustrans

Sustrans are deleting some routes from the National Cycle Network map and disowning others, as seen on www.theguardian.com/travel/2020/jul/19/national-cycle-network-sustrans-cuts-quarter-uk-routes-safety-grounds and on the mapping on www.sustrans.org.uk

The routes sound likely to remain signed on the ground for some time and slowly fail, possibly having their route numbers covered up.

How should these be remapped? Should they be listed as soon as possible (as soon as known from a clean source, rather than Sustrans's Ordnance-Survey-derived mapping) as "abandoned:network=ncn" and "abandoned:ref=..."? Should they be converted into "network=lcn" routes with named from/to places? MJ Ray (talk) 13:22, 20 July 2020 (UTC)

My feeling is that this is the point when we really need to agree what should be mapped and what not mapped. Historically Sustrans mapping has been used as something of a guide, even when routes don't really exist on the ground. I'd propose that OSM should always record what's on the ground, not what Sustrans pretends exists, and we certainly shouldn't remove routes that actually are still signed just because Sustrans hasn't got around to removing signs (this might take many years). I'd like to see some agreed way to record something like "signed=yes | partial | no" (which gets a bit complex because we use relations, but could be managed I think. Rostranimin (talk) 21:42, 15 August 2020 (UTC)
If being mapped means it's signed then we have adding another value to state (besides proposed and alternate - taginfo says abandoned and disused are already used a little) and the Lifecycle_prefix tags as consistent options. MJ Ray (talk) 10:33, 17 August 2020 (UTC)
I have a wider purpose in suggesting the use of a tag to indicate signage state. As I raised here, I see a growing problem with the inclusion in OSM data of unsigned routes. Some of these are relatively formal things - with some kind of formal description. Others are just someone's hobby route - perhaps described as part of a web page or something. Practically I think that the easiest way to deal with these isn't to argue with people over their inclusion, but to mark that they don't exist on the ground - that they aren't signed. What I proposed on that discussion was...
  • signed=yes (route is signed using an identifier and overall signage is sufficient so route can be followed without a map)
  • signed=frequently (route overall can mostly be followed without a map, but there are some problems)
  • signed=rarely (route signs exist, but can't be followed without a map)
  • signed=no (this is an unsigned route)
Rostranimin (talk) 18:00, 17 August 2020 (UTC)
Well, I think hobby routes have no place on the map. They do not really do anyone any good and would probably be better as GPX overlays/downloads on the hobby website - unless they've got signs up somehow. There's one near me which I probably should remove from the map: relation 3773003 doesn't exist on the ground or even on the linked website any more, as far as I can tell. A more difficult example is the LEL route 8885353.
I think it's OK for the map to contain a route that is unsigned but expected to be signed eventually (state=proposed). A more difficult question is whether it should show routes that are unsigned because it's not good enough for inclusion in the network and the operator hopes to build something better but this is the best completion for now (state=recommended).
I think you're tackling a different problem with the signed=* tags: a signage description, similar to the surface descriptions, might be useful, but isn't quite the same as whether a route is still maintained/monitored and has an operator/sponsor.
MJ Ray (talk) 23:15, 18 August 2020 (UTC)