How appropriate… Here comes an actual discussion on the topic:
Did you note how often you and others now use the word ‘route’ in these cases? It kind of converges with my thoughts from the threads you started last year: actually, a path is a kind of route and not the reverse.
Consider this: all physical ways are aimed at allowing people to move from one place to another, and they have in common to be visible, thus allowing users to decide where to go. They are a particular kind of routable objects, others being e.g. free routing across a pedestrian areas, series of cairns in a rocky area, skitouring slopes between two passes, a ferry route, etc. All routable, none except ‘highways’ visible as a linear trace on the ground.
Therefore, rather than reusing the highway primary keyword and adding a secondary keyword to state that actually it is not a highway of any kind, I would say that we need a new primary keyword that encompasses highways as special cases… and this thread suggests that route might be a good candidate.
See: we already have various special cases handled as route=* , and it does not break any existing semantics. We could have route=cairns, route=free (with just two points), route=scramble, route=climbing, etc. Generalizing the use of route would rather consolidate the semantics of the whole. We would just have to imagine that all highways have an implicit route=highway tag. And the same goes for relations: we just have to say that routes can be made of routes, which is already true actually.
6 Likes
That’s an interesting thought! The first objection will be that it breaks the good practice of Mapping what’s on the ground. However I think it’s a good practice only because it ensures that what we map is Verifiable, but it’s just one way to ensure it. Surely there are other ways to ensure that something we map is verifiable. So if we want to map a route that’s not visible on the ground, there should be another way to verify it. If it can’t be verified, it shouldn’t be mapped.
2 Likes
Yes, see for instance administrative borderlines.
Actually we already do that for hiking routes : they can be verified through signage on the ground or from official sources (e.g. routes E1 to E12). Another way of verifying them, but I mention it as a philosophical thought, is through Strava and equivalent services.
In any case, the same question can be raised whatever the tagging method used. I see that as two chained questions: should we map those routes, and how.
Unless a strong objection emerges, that is actually a proposal I would probably vote for.
Though, up until now, is not route reserved more or less for relations? (like route=hiking)?
Also there is already trailblaized:cairns so in your example, that should fall under :route_free if the cairns are the only thing there is.
Mixing route markings with the mode of travel for the route is a terrible idea.
Aside from that, you can pretty much already do that. Just map the cairns and add the nodes to the route relation without ways.
But: route relations should be reserved for actual verifiable routes (i.e. signed or otherwise common knowledge)
There are nearly 30 thousand ways tagged as route=ferry, among other examples.
I’m not sure that many people know that. Last time I checked, route relations with only nodes did not seem to be well-documented. Do they get rendered anywhere? Given that the motivation for using highway=path seems to be to see the route on a map, it seems unlikely they would change their practice unless there is a prospect of some practical support for another approach.
Then the effort should focus on actually improving the tools and renderings instead of trying to misuse something else just because it looks prettier.
Who suggested that?
You in the section that I quoted.
OK. Let’s try to understand with the these new details that you shared. If I was to reformulate your statement into something that contributes to the debate in a constructive way, would something like the following reflect your thoughts?
If we were to tag routes that are not actual paths with another tag, we must be careful to find values that as much as possible capture how the route can be observed (cairns, free, etc) and not how the route is traveled (scramble, ferry) .
Close enough? If yes, could you elaborate on the reasons so that we all can continue to contribute to the debate?
1 Like
I think this is the key. But not in the way you probably meant it.
For example, adding a route that is marked with an occasional cairn presents exactly what’s on the ground. That is the whole point of mapping it. The issue mapping is trying to solve, is that it may be hard to find the path/route on the ground due to various conditions.
To make it even simpler, I don’t see how mapping a path between two points is breaking the “on the ground” truth. It is certainly possible to walk between those points, so what exactly breaks that rule? It is the “on-the-ground” truth.
Edit: I definitely support another (primary) tag for this purpose, as I’ve mentioned in these discussions. If the Path is reserved for something that is clearly visible on the ground, then we need another tag for something that may not be clearly visible to everyone at all times. For example, I was following what looked like the best route for me and I ran into a series of cairns. This means that it was also the best route for someone else, who was kind enough to put a lot of time and effort to collect fairly big stones and arrange them into cairns along the route. So, the path/route/way/reasonable possibility is there.
Probably finding the correct name is the only problem. Hope someone comes up with a proposal on this!
Thumbs up not for deleting what was there 
Recently, when examining ways both tagged mtb:scale and sac_scale to learn, how theses two scales relate or not relate, I stumbled over highway=pathless - A very fine example of AnyTagYouLike – verifiably there is no path there – while the term itself features prominently in route descriptions. Of course the tag later was deleted, because it is not part of some blessed ontology. I’d never have deleted this tag, I find it quite plausible. I also do not see any chance of getting a proposal accepted for this tag though 
To be clear, I deleted my post because a substantial edit was made to the prior post while I was typing, and I didn’t have a chance to rewrite mine to reflect that.
It is worth noting that this is not the first time a similar topic has been discussed. I haven’t really read these other threads yet and there may be more, but I’ll just leave these links here:
3 Likes