Tag trail_visibility: Proposed Improvements for this Descriptive Tag

[erutan]

@rhhs

What do you think of the revised version? Is it better? Can it be improved?

This is better, and essentially describes a path_visibility replacing the original intention of what I’d call trail_surface_visibility.

The key is not about route visibility (how easy it is to choose the right trail at a trail junction)

Route visibiliity sounds strange - I would change it to routing (and mention the more formal term route relations). Pathless paths are often called routes as a route is just the specific path you are doing or want to do regardless of any physical conditions around it, where routing is trying to get from one place to another. This really has very little to do with the trail, and more to do with junctions & signposts as the parenthesis makes clear. It could be worth adding a link to

Path sometimes invisible, route partly pathless

Here you are using the definition of route I am referring to - unless this somewhat means that routing between paths is mostly pathless? That doesn’t make any sense to me. Using the word route does not automatically make something about routing, which seems to confuse people that are used to thinking about general map routing.

trail_visibility=no can be useful to indicate that a route is passable, but no path is visible.

This is awful. I can literally make hundreds of routes that contain “passable” terrain to the same location - do we really want OSM full of paths that are just “passable terrain”? See Tag trail_visibility: Proposed Improvements for this Descriptive Tag - #80 by erutan

Is there a formal movement away from using “ground truth” to map paths? I had thought this was a core tent of OSM, but perhaps I’m wrong.

If the overall path visibility is “no” then it should be a node network of digital “markers” to keep people on key points or decisions. IMO it is no longer a path, by definition it is “pathless”. If someone can’t routefind on their own they shouldn’t just be staring at a random person’s GPX track creating new trails by packing use on a concentrated area. If nothing else they will never be accurate enough to truly routefind in semi-technical terrain (even ± 3 meters can get you in trouble). I guess we could make a highway=pathless lol

@rhhs

So trailblazed:visibility only really matters when he has to make a choice between different paths, i.e. at junctions. There it could be useful info (to know how much attention he needs to pay, to know how much time he might need to spend on finding trail markers and on walking down the wrong path, etc.).

@rhhs

Here I don’t think it makes any practical sense to separate markers from any other visual clues: anything that helps to follow the mapped trail (the thin dashed line on the map is what I mean by “path” or “trail”) should be taken into account.

Doesn’t the original definition of trail_visibility (which was accepted at the time) also help when trying to find an intersection? If you have no markers but an excellent clear trail surface junctions are straightforward, if you have no trail surface but a lot of markers and a larger marker at a junction (plus further markers heading in more directions) junctions are straightforward. Is there a clear argument as to why only the latter is important and the first case isn’t?

In terms of pure routing, overall path visibility is what matters. Trail surface, marker visibility, and overall path visibility don’t always add up - I gave a recent example where surface visibility is horrible, markers are intermediate, that sort of averages out to the overall experience of the path visibility is bad. I can honestly see an argument for just simplifying things and relying on the new trail_visbility and trailblazed without the related :visibility but at the point where people are arguing that trailblazed:visibility is important I can’t see why trail surface is somehow irrelevant.

@osmuser63783

I have nothing against breaking down trail_visibility further into trailblazed:visibility and trail_surface_visibility. It could be useful for renderers, they might want to show an indistinct but well-signposted path differently from a well-worn but unmarked one. It’s just that I can’t see myself using trail_surface_visibility much: we’re debating edge cases here but for the vast majority of paths I imagine the value will be identical to that of trail_visibility.

It’ll come down to a regional thing, as well as how often people are hiking on abandoned or informal trails.

In areas of heavy vegetation, dirt, or sand you will tend to have paths that express themselves by solely trail_surface_visibility unless you are briefly crossing a streambed or something (wet or dry) in which case there will be markers. In areas of long stone slabs (alpine granite, desert slick rock / sandstone) or consistent snow/ice coverage paths will tend to be driven by markers. In the past few weeks I’ve hiked on areas that were excellent on one and no on the other, and vice versa.

One interesting case are trails that have a good surface visibility, and are occasionally marked where surface visibility is poor so the overall visibility of the path is still excellent. In that case I think just having trailblazed=cairn, trail_surface_visibility=good, and trail_visibility=excellent would tell the story pretty clearly - you will lose the trail at times, but if you look for a marker you can find yourself back on the path again.

In the past few weeks I’ve also hiked on multiple pathless paths that I created from my own routefinding, but that I think are not appropriate for OSM or any general use mapping. I shared some screenshots of my personal waypoints with a local that enjoys off-trail, and he may share them with people he knows and trusts, etc and that’s fine but I wouldn’t put them up as a “path”. His “route” will be slightly different from the one I had taken as we have difference preferences for terrain, directness, exposure, etc… that’s part of the fun IMO.

@yvecai

I wouldn’t throw aerial imagery in the discussion about on the ground visibility.

I just deleted a path (after reaching out to the creator) that looked like it could be a path using imagery (it followed a wash and some game trails) but in reality was no stronger or weaker than many areas around it. It also violated the LNT off-trail ethos of the area it was in as some of those game trails that were more direct had living biological crust around hoof prints, so I ignored it and kept to washes, but was curious if it actually had some historical basis. It didn’t.

That said tracing old jeep roads or tails based on imagery vs relying on a bunch of smoothed GPX tracks or old USGS lines makes sense.