|
Public Rights of Way in eastern Hampshire (UK) are being tagged as access=private (the mapper states this is because they have been established by easement). This results in ways on displayed maps losing their identity as footpaths or bridleways and being indistinguishable from genuine private ways. This means that the map has become useless for hikers, cyclists and other users of PROW in the affected area. If this continues OSM will suffer serious (and probably irreversible) damage to its reputation in the UK.
This question is marked "community wiki".
showing 5 of 8
show 3 more comments
|
|
We can broadly distinguish three different situations when adding access tags to highways which are also Public Rights of Way (PRoW). Firstly, Normally certain OSM highway tags can be presumed to imply various access tags, roughly along this lines:
The implied defaults may vary by individual countries but are usually fairly consistently mapped within a single jurisdiction. In addition in England and Wales designation=public_bridleway emphatically implies foot=yes, bicycle=yes, horse=yes. Secondly, use of the highway=path tag does require that access rights be made explicit, even for foot=yes. In general I would only tag a path with no known access rights with access=private, but for those with access rights specific each one. Thirdly, use of a generic access tag is appropriate when the highway tag is not one of those typical for a PRoW (e.g., a driveway to a house, a farm access road, or an agricultural track). It seems in this case there is a confusion: the land over which the highway runs is private. Usually the formal written designation of a public right of way will include the notional width of the relevant path. Here is an example where the PRoW as specified is narrower than its apparent extent and the landowner has requested that people stick to the path: photo Thus there are three potential situations depending on which highway tag has been used to map the way. The addition of an access=private to the first and second categories mentioned above rather complicates the tagging for no apparent additional informational value. As the OP states it has the consequence of negating any use of the standard CartoCSS for planning walks in this part of England (although recent style changes have made it less useful for this anyway). So broadly speaking I feel this is over zealous tagging. It is not wrong, but like many such tagging schemes has the effect of making the data harder to use. Furthermore it is highly divergent from the usual way of representing such cases in Great Britain (85000 ways vs. 1300 (August 2014 data)). Normally such a major change from the existing consensus should be discussed up front on talk-gb. This usage is not generally suitable for Public Rights of Way in England and Wales for the reasons stated above. OpenStreetMap tagging is mainly not about providing a legalistic precise view of the world, but a pragmatically usable one ('perfect is the enemy of the good enough'). I'm concerned that that this way has been subject to an edit war. Edit wars do not reflect well on any of the parties involved. Please consider your mapping practices before making any more changes of this sort. 2
The wiki has a page about implicit default access values. It also states that there are implicit default access values for highway=path except that for the United Kingdom explicit values are required.
(20 Oct '14, 19:50)
scai ♦
Yes, useful link, which confirms what I say here, although I suspect most GB/UK mappers dont read the wiki for this info.
(20 Oct '14, 20:00)
SK53 ♦
|
|
Wow - way 183471443 does have a lot of tags on it! I'm not entirely sure what " It's mapped as a " I don't think that adding " Obviously the above is my personal opinion - a glance at the state of footpath tagging in the UK will reveal that's it's not the only one! ** I know that in places like Holland young kids get to ride hairdriers-with-wheels on footpaths and cycleways, but we can leave them out of the equation for the UK. 1
... and one thing that I forgot to mention. I notice that just down the road there is this track with the name "Wayfarers Walk, Three Castles Path". Unless that really is the name of that piece of track I wouldn't have added that as a name but instead added it to the relations that have those names, which has already happened. These will show on sites such as this one.
(20 Oct '14, 18:23)
SomeoneElse ♦
|
Can you give an example of an affected way? It would be interesting to know whether it's the current tagging or the rendering by the openstreetmap-carto style that is "incorrect" (for some value of "incorrect").
Recent changes to the standard stylesheet such as this one have made the "standard" map less useful to me (as in, not useful at all) for displaying rights of way, but it's entirely possible that a way that isn't shown by the openstreetmap-carto style is tagged correctly, and the style changes were made for valid reasons, after discussion on Github
@Sailor Steve: I agree with your concern and would say that this is sufficiently against common UK tagging practice that it deserves discussion. But... please try and avoid hyperbole like "probably irreversible damage". Thanks!
SomeoneElse - Example 183471443 (north of Alresford) - part of Wayfarers Walk & Three Castles Path.
Richard - Thanks for your reply. However having worked many years in a service industry I know reputation is easily lost and difficult to restore.
Independently of the wrong access tagging already mentioned by SomeoneElse: Your question seems to be mainly about the rendering of access=private while other access keys are present. This is indeed questionable but should be discussed at https://github.com/gravitystorm/openstreetmap-carto/ (where the rendering stylesheet is hosted) instead.
Thanks everyone for a lot of useful info in your replies.
My point regarding the use of the access=private tag is that it renders a PROW as a private way, this is not a problem with map rendering - it takes this tag (which is the top level) as the rendering priority, which is as it should be. If access=yes and motor_vehicle=private are applied instead the way will render correctly and routing software should recognise the way as not available for planning a vehicle journey on public roads.
The way I quoted, although designated as a bridleway has parts that are a quite narrow path, perhaps quite challenging for a horse rider, so the path tag is not inappropriate. I agree that many of the long list of tags are superfluous, this seems to be an unavoidable fact of life on OSM.
The use of access=private on PROW's appears to directly contradict its OSM wiki entry. I have made these points to the mapper through OSM messages, but he seems unable to see the logic. Most mappers make mistakes, I'm sure I have made more than my fair share along the way, but that's how we learn on the job! But how to deal with a situation like this where the map is being damaged? That's where you guys come in.
I don't see how "The use of access=private on PROW's appears to directly contradict its OSM wiki entry". An England and Wales public bridleway by its very definition allows certain kinds of access - it doesn't allow all and only disallow some. If you used your suggested "access=yes; motor_vehicle=private" combination on a farm track that was a public bridleway it would incorrectly suggest that horse-drawn vehicles could use it (they legally can use "restricted byways", but not "public bridleways").
See also the discussion on https://github.com/gravitystorm/openstreetmap-carto/issues/765 .
I used "motor_vehicle=private" as an example, where a private way has some access rights each barred access type or group should be included as a *=private tag. If all vehicles are barred then vehicles=private would be used.
Both ways are equally valid. It doesn't matter if you apply "access=private" and allow individual transportation modes, e.g. via "foot=yes", or if you don't apply a global access condition but disallow specific transportation modes, e.g. via "motor_vehicle=private". If used correctly, both solutions will represent the same conditions and both solutions will be handled equally by a decent router. In contrast, these two solution will get rendered differently on our main map because it is somewhat impossible to have a meaningful rendering for each individual access combination.