I put it there (I always meant golf_cart for the sake of comleteness without digging into it much). I misread the wiki, I did not notice that they are actually already using highway (though almost evenly split between path and service). It should definitely be a subtype, it already is, and this does apply:
(ideally, they have no overlap with the new tags):
TLDR: they should stay golf-carts. Sorting out the service/path split is out of scope for this proposal I think (to me they should be their own first class tag).
Thereās a lot of backstory behind this split. To oversimplify: there were a couple rounds of miscommunication when introducing two different presets, one for golf carts and another for maintenance roads that golf carts could also use. Both wound up including highway=service by accident. Before we could sort it out, a golfing game developer and an outside community of golf course gamers latched onto these presets and mapped a ton of them.
There was a proposal for a dedicated highway=cart_path or somesuch, but it stalled over concerns that the paths would go missing from OSM Carto and other renderers. In the meantime, these āservice roadsā were likely causing real-world accidents, so as an emergency stopgap, the preset was changed over to highway=pathgolf_cart=yes. (We decided against golf_cart=designated because some cart paths are on private property.) Maintenance roads continue to be highway=service. This change didnāt come with a concerted effort to retag existing paths. We just assumed the golf mappers would automatically update the paths to match the presets, which did happen to a significant degree.
If only there were a similarly herdlike bunch of mappers for paths in general. Then we wouldnāt have to worry about what to do with all the existing paths.
I still believe that it would be great if we could avoid exposing new contributors to path=pathless. Can we add other candidates such as trail and route for the name of the main tag?
Do I understand correctly that this is proposed to be a secondary tag to go with highway=path? āharder than sac_scale=diffficult_alpine_hikingā means beyond the SAC scale. For the average pedestrian, they āneed to be circumspect about what they are doingā at much easier sac_scale values, say sac_scale=demanding_mountain_hiking and beyond. We could decide that these paths should additionally be tagged with path=mountaineering. But wouldnāt it be even better to tag them with the actual SAC value, which is an already established tag which carries more detailed information than path=mountaineering. Why introduce path=mountaineering, then?
Most paths currently tagged with trail_visibility=no are not dangerous at all, so tagging them with pathless instead as an indication of āthink twice before going thereā is overkill. What is the difference between highway=path + trail_visibility=no and highway=path + path=pathless? Donāt they both say the same thing? And wouldnāt a path with path=traces be better tagged with trail_visibility=bad or trail_visibility=horrible (established tag that offers more detail)?
This tries to address the first problem I mentioned. I have some ideas about the second problem and hope to have some time to write them down and start a thread.
To my surprise a recent poll in the Austrian category regarding tags created by a helper plug-in, whether to use highway=cycleway or highway=path as base for shared-use foot/cycle-ways as of now stands at 7:1 in favour of cycleway.
Mind you, shared use is one of the mainstay of path. I wonder if path would have ever been proposed if it was not for shared use. I bet that relentless spamming of the forum in recent months played a role in this poll results.
How about? RfC: Deprecate use of highway=path for municipally maintained infrastructure. Use cycleway/footway there, and only there.
Single out the other end of the spectrum: But I fear, that will meet exactly the same strong opposition. A bit less ambitious, how about proposing for a start: Do not use cycleway for mtb-downhill-tracks, do not use footway for backwoods hiking or social trails - use path there; āand only thereā can come later.
No. Please read the proposal again. That is explicitly mentioned.
My opinion is that paths that are largely not visible on the ground should not share a primary tag with paths that actually exist. I am never really happy when I try to follow a path to realize that it does not exist. I mostly hike in the mountains, there your observation does not hold. I guess it is different in the UK (everything is different there?). I would be happy with having two primary tags for physically non existant and difficult paths.
As for your suggestion, I think it is not really a solution. I think path is broken and even having absolute coverage of secondary tags would not solve its problems. Which is not saying more secondary tags would not be a good thing. It is just that āpathā is too wide.
Yes, please. If you have additional suggestions, spill them out. Also, the posts are wiki entries so feel free to add stuff. I see this as a group effort. Hope something comes out of it.
Being careful with the value of the tag is one way of doing it. But I forgot to mention that the value path might help up a lot with the transition. Using pathway=path or trail=path or route=path would provide a smooth transition from highway=path. For this, we would need to be careful with the name of the tag, not only its values.
Except if we donāt call them paths. We want to represent a collection of trajectories that can be identified through different visual clues on the ground and used to go from one point to another. Paths are a subset in this collection.
I know that when I start a hiking route I will be using paths, tracks, traces, climbs, etc. Itās consistent in my eye.
We have Tag:path=mtb - OpenStreetMap Wiki in use currently. To me the two values here mtb_single_track, downhill_trail are confusing, why not just mtb?
A predominantly downhill trail should be indicated with an Key:incline - OpenStreetMap Wiki value of down, and the uphill/downhill aspect should not be within the path key value.
There are multiple flavors of mountain biking, but a very popular one is a downhill descent over a specially designed trail ā mtb_single_track. Obviously, cyclists have to return to the top somehow, but they will not do that up the same trail, but over a common forest path, or even using public transport where available. Thus, āuphillā MTB paths are in 99.9% cases just ordinary paths, and as a hiker you donāt have to worry about that much, since cyclists going uphill will be slow and harmless. Thus the proposal to separate only mtb_single_track, which also has special construction features.
I would propose mtb_downhill as the tag value, since it is obvious to both MTB enthusiasts and general public what is meant.
I share many of the general reservations already expressed by others above (and ones that I expressed in the previous discussions). But thereās much that I like! Furthermore, the proposal could at least form a good and specific working path forward (as it is also obvious that e.g. merely adding more secondary tags to paths wonāt fly).
One specific comment: I find it strange that shared_use is filed under the heading of āSpecializedā paths. Isnāt it just a karrenweg with an additional motor_vehicle=no tag? Other ways under the same rubric contain stuff like mtb_single_track and mtb_downhill that really are highly specialized path types and will for that reason see very limited use. shared_use, on the other hand is in my understanding the very default value and use-case of all=paths. Ceterum Censeo, @Hungerburg nailed much of my feelings in their comment above, touching also on this specific issue.
I would tend to agree. However, we are using, pardon - relentlessly spamming, the forum to discuss if a separate value makes sense or not.
Many other items are segregated - types of buildings, roads, sports grounds. Maybe it would make sense to also segregate types of paths.
Oh, okay! In that case Iād agree (as I said) with @Hungerburg above and say that we should encourage people to tag ways with those kinds of traffic signs with highway=cycleway (and additional =designated and segrgated= tags) instead of =path.
However, in any case shared_use (and/or karrenwegwith the motor vehicle access-tag)āin their meaning posted in the proposalāare widespread and ubiquitous, at least in my neck of the woods. Therefore not āspecializedā in any very meaningful sense.
Taking the risk of relentlessly furthering the spamming of the forum with an unstructured stream of consciousness, Iād like to make an attempt at making our criteria explicit when deciding to segregate or not:
because it makes a difference to end users
because data consumers will know what to do with them (more or as well as if not segregated)
because it will make things easier for mappers (or at least not more difficult than if not segregated)
āSpecializedā here also means āmore to the pointā in a way. The first four categories are brand new, while these āspecializedā (for a lack of a better term) are more specific (maybe I use that term, instead) and defined/verifiable.
However, they also overlap as the first distinction is based more on physical characteristics while the second one is based mainly on function or legal designation.
In the end, I think the stereotypes for paths should be ones that are commonly used and recognized, not defined with mathematical precision and definitions. Any such distinction will be practically useless. Thatās exactly why many think that having 20 parameters is enough, because they can describe any type of path that may exist. And that is technically true but practically totally useless (and, hence, mainly an unused feature).
forward (as it is also obvious that e.g. merely adding more secondary tags to 