Should we strive for a global or regional consensus for things like trail visibility & difficulty (sac etc), and possible pathless paths?

The pictures might not be perfect, but they serve the purpose:

T2 means, you have to lift your legs and watch your step.


Nah, it was written by people who liked to draw paths to the standard view and needed a cover-up, Re: [Talk-de] Wanderwege, Klettersteige, Kletterfelsen so they could not be blamed, if somebody went astray or got hurt. Certainly not mountaineers.

Mountaineering only came in by choosing the SAC hiking scale, which was created for approach paths to the huts, where mountaineering starts and routes to some easy non-technical summits.

1 Like

Not sure what that means. The reply to that was someone talking about their mountain tour (whatever that means).

Regardless of who actually wrote the wiki, the content is heavily aimed at people with mountaineering experience. One of the T5 examples has a via ferrata cable. I would treat that similarly to a rope in terms of protection and why it exists - if you are dragging yourself up a cable it’s not scrambling heh. I tended to ignore the simple stuff in Canada, NZ, Chile, Argentina etc that have a more European influence in putting iron all over the place. The second T6 photo shows a nice ledge near the person doing it but given that it overlaps into UAII 2 I doubt the entire route is like that. I could see someone inexperienced wanting to be roped in there though it wouldn’t be done as a technical climb, though I guess there’s enough via ferrata routes for people like that.

How many non-mountaineers take T6 trails to huts? Is that a common thing people casually do?

What? Are you seriously saying that tagging a way with a certain trail_visibility in OSM implies a certain sac_scale in OSM?

The keys are independent in OSM.

As for “ignoring all the hard stuff” - again no. Both OSM keys have a range of values, so most people, me included, will simply choose the best fit. If you don’t want to use a particular tag, you don’t have to.

Well, you are free to take the a-historic PoV. Historically though, TV was “approved” in the same proposal as TX(1…6). That is why both share the same picture to show what excellent TV1 is and what hiking T1 is.

I happen to repeat, this split has not been devised by mountaineers. You cannot do T6 hiking in the hall, like you can do 9+ climbing in the hall. Orienteering is part of the game!

PS: My home town is currently home of the world trail running championship 2023 - it goes mostly over T1 T2 paths Look at for video. I posted mountain_hiking trail created in the old town picture in this thread already.

1 Like

Yes, that is clearly what I was saying.

Here’s another take - if we’re going to show pictures of path difficulty it’d be nice if, say, more than 1 out of 9 photos actually showed something resembling what people consider a path if we’re aiming for clarity? If the current pictures are the clarity we’re aiming for, then I think that says a lot.

Sure T2 has a paint marking on it, and T3 is fine to show pathless (see trail_visibility=intermediate) but they’re not great examples of average hiking paths that require you to “pick your foot up” or “maybe use your hands”. It goes straight from a picture of a single track trail to routefinding your way on talus with nothing in between. Even for trails that fall into those technique levels those aren’t the average use cases, and those are the only pictured examples.

It doesn’t help that there’s text at the top of the page that basically says don’t read any of the visibility stuff here but then the page still has visibility language scattered all over it, and has for a very long time. I bet there have been people applying the visibility language in the values, because they went there to read the values and just skimmed the rest. Why not take it out? What purpose does it serve?

The best fit isn’t the best fit. One key is for 64% of trails, two more account for 35%, and that’s in a place which is more likely than most to have values for those other two keys. It’s definitely not the only place, and you can really cram them in to make them work, but probably 80-90% of trails are in the first two values and we’d be better served looking at other systems - T1 is basically the first 3-4 easiest grades of 5 trails in the US, first 2 in Italy, first 2-4 in Australia & Tasmania etc.

Sure we can just use T1 or T2 for the majority of trails, ignore the details, selectively pick aspects of T3 and T4 when they apply (which is controversial). It works, but it’s messy and subpar for general use. There’s a reason many other countries are more different to SAC than similar to it - the system was devised by an alpine club. Can you explain why SAC better serves the needs of OSM users than those other systems, or this proposal other than “that’s the way it is”.

I’m aware that trail_visibility is its own key, but it’s basically the visibility portion of SAC ripped out. The example photo of good where the trail doesn’t exist is likely from a T2 trail, therefore someone that knows SAC would label it as good because it maps to that system. It’s the same unbalanced distribution of focus and why there’s one good, one intermediate, four bad, and no actual excellent.

You really are overthinking this :slight_smile:

1 Like

From what I observe in the are of my local knowledge, you are right though: trail_visibility is more commonly tagged on informal=yes paths than on managed hiking paths. Likely, because the hiking paths all are TV=excellent (according to wiki documentation) or good (according to tagging practice.)

1 Like

I understand that this discrepancy between practice and documentation hurts.

There’s many people unhappy with trail_visibility values. There’s less people unhappy with sac_scale, but it’s still not a well understood system outside of continental Europe, it doesn’t always apply to local conditions very well, and while it’s not described well it’s set in stone. Heck, many other European countries don’t think it’s a “best fit” for them - whether formal state entities or alpine clubs their systems have a very different focus. Anything can be made to fit anything, but that doesn’t mean it’s not counterproductive or suboptimal to do so.

That catching up will be slow.

A problem is that we don’t have a good way of getting data about actual practice. Some of the things in the documentation of sac_scale and trail_visibility don’t make much sense when you take them too literally, so I suspect that people don’t take them that literally. But I can’t prove that suspicion. Looking at the pictures for trail_visibility=good, I also suspect that more people would tag the first picture as intermediate than good, but I have no easy way of checking the actual mapping practice. This is different from roads, where you can identify examples with Overpass and then check them on Mapillary. So we all only know how the tags are actually used on the paths we’ve hiked ourselves…

1 Like

Let’s try another approach.

Forget about the OSM wiki - pretend that it does not exist. Of the two concepts that we’re thinking about here, let’s pick “trail_visibility” as an example (it doesn’t matter which both are tags which have values on a scale between “very very good” and “very very bad”.

Now think about the paths that you might want to tag with trail_visibility. You can probably put them into an order yourself, from best to worst. So far so good, but you still need to know which absolute values to use.

To determine that, look at other trails that have already been tagged with trail_visibility. How would they fit into the list of trails that you want to categorise? Once you’ve done that, you now have a series of absolute values that you can apply to the trails that you want to categorise. Job done.


I agree and this is why I suggested a gallery like Key:smoothness/Gallery - OpenStreetMap Wiki for both keys

Ideally I’d love to see something like Komoot Trail View but free and integrated with OSM. A picture is worth a thousand tags

Survey (or the recollection of a previous visit) is generally the best way to do this. If I remember correctly, the change to Lord’s Rake in the English Lake District followed a talk-gb or #OSM-GB discussion between people who’d been there before.

That’s recursive though. Current usage is influenced off of the wiki, guidance in the editors, initial editing / data imports etc. Best case scenario is you still have more bad data than you normally would just through noise and subjective takes as you have people trying to follow docs and people ignoring it. Multiple people have argued that altering the trail visibility values is impossible because that would break past values - either they exist in a vacuum or they don’t.

There’s a meaningful difference between what most countries would consider excellent and the OSM excellent. The US has 2 values above it, Italy 1, Australia 2 I think, Tasmania 3-4. Austria & Switzerland are more in line with the current system, but even some Austrian users feel it’s not a good fit. :stuck_out_tongue:

I remember @Hungerburg saying something about an editor that conflated sac and trail vis unless I’m mistaken.

Interestingly, trail_visibility is noticeably higher at 29 902. While that’s still a small part of the global 667 755 usage the proportional difference seems to indicate that more people in the US are comfortable with trail_visibility than sac_scale. Which makes sense. You think they’d be tagged together in most instances.

3.8% of the global use of sac_scale
4.5% of the global use of trail_visibility

It could also be that tens of thousands of European trails don’t have TV because people there assume that it is baked into sac_scale.

Sorry, I don’t understand - “excellent” is just a word in the English language that happens to be used to correspond to the “very best” end of the scale used in OSM. Instead of “excellent” it could be “frabjous” - it doesn’t really matter. Perhaps are you saying tthat somewhere in the USA there are places with trail_visibility=most_excellent or similar? If so, taginfo doesn’t find anything obvious.

If you’re saying that there’s a trail catagorisation scheme in use outside OSM in the USA that has most_excellent (or whatever) in it that’s not really relevant - it’s how we tag things in OSM that’s relevant here.

There’s a LOT of granularity being lost at the best at the expense of the worst in how things are tagged in OSM. The very worst are routes that aren’t even supposed to show up on maps (SAC T5-6 in Austria, AWTGS 6, PWS R, etc). Some systems run out before hitting the “worst” bounds of that but that is what our current trail_visibility is built around.

OSM excellent covers NFS 2-5 (top 4 grades), CAI T & E, PWS W1-T3 (top 5 grades), AWTGS 1-3/4? and SAC T1

OSM good covers NFS 2, CAI EE, AWTGS 4, PWS T4, SAC T2

OSM intermediate covers NFS 1-2, CAI EE, AWTGS 5, PWS T4 and SAC T3

OSM bad/horrible covers NFS 1, AWTGS 5, PWS T4-T5, SAC T4-5

OSM no covers SAC T6, AWTGS 6, PWS R.

At the very least the CAI T / NFS 4-5 = actual excellent, CAI E / NFS 3 = actual good makes a lot of sense. If you have to search for the trail, it has large gaps, it doesn’t need to visible if routefinding is obvious etc OSM good isn’t really “good” to your average person. It’s worth noting that these cover aspects of trail building that don’t have much to do with visibility (but also many that do).

I’m not sure how a path which is mostly visible isn’t sometimes pathless but a path that is sometimes invisible is pathless, so I imagine there’s some fluidity in those unless you really pay attention to requirements, which we’re apparently supposed to ignore. I treated OMS intermediate gentler than it probably deserved.

I’m assuming that you are talking about “trail_visibility” in OSM here.

The OSM trail_visibility tag is defined within OSM. You can either read the wiki page or look at how existing ways with a similar trail_visibility have already been tagged (I’m assuming that that alphabet soup is all international track grading systems). Where it came from is irrelevant.

It’s certainly possible that a some hiking club somewhere will have more values for their equivalent of “trail_visibility” than we do for for ours, and yes, you could argue there that “granularity is being lost”

I’m assuming that you are talking about “sac_scale” in OSM here.

SAC T5 and T6 have a direct analogue in OSM tagging - sac_scale=demanding_alpine_hiking and sac_scale=difficult_alpine_hiking. Whether a particular OSM way shows up on a particular OSM map depends on the person that created that map. Many people exclude OSM ways with those sac_scale values from general purpose maps.

If so, the current scheme is lacking a value for this, I’d see there the gold-standard hiking path according to the Austrian classification:

This is not what excellent is in OSM - which looks like this:

When looking at the thumbnail, I am not certain, if I see a path where the cut branches are.

ADD: There is quite a discrepancy in how values are spent between Austria and the U.S.A - From what I observe, people here in Austria do not tag according to the wiki.

1 Like