I agree on the second quote more or less (though I’ve met a people that are good technical gym climbers that have a hard time doing basic routefinding in wilderness), but it’s certainly possible to have a “simple” trail in terms of technique that is hardly visible. You gave an example of a path which went on from a bridge over a grassy hillside that was hard to see or follow despite being T1 in that it wouldn’t be “harder” than just walking in an urban park.
Sometimes cutlines follow power transmission lines over fairly long distances (kilometers, tens of kilometers) and have highway=track of a grade (tracktype=grade3 or so) to support a high-clearance and/or 4-wheel-drive maintenance vehicle for access to power towers. These can often be seen in satellite imagery, if they exist.
I have mapped many man_made=cutlines. They are cutlines. I would not “automatically” consider them a path, even though it is very likely that they might be or are a path. In the latter case, I wouldn’t change any tagging in OSM, even if in the real world I were to use the cutline as a path (because I happened to be hiking in the area and found myself on a cutline and it was a possible route for me to hike to another place I might want to hike to). If I then later fired up an OSM editing session, saw (or entered) the cutline as data in OSM and thought to myself “hm, should I add highway=path here?” I would only do so if there were a clear highway=path (or highway=track) ON that cutline (in my personal experience / from my earlier on-the-ground survey). Especially if I didn’t know there were a highway=path or highway=track there, I wouldn’t enter one by speculation, simply “because it is a cutline and is likely a path.”
I think these attitudes and behaviors are in harmony with “how we map” in OSM.
1 Like
Amusingly the first photo for “trail_visibility=good” is labeled with the following:
“While the way is not directly visible. As soon as one advances around 20m, way goes up on the right and is very well visible again. Only feature is that for a few (50) meters the trail is not really defined. Orientation is very easy however.”
That photo always confused me, it makes sense with the caption.
The second is better as a faint path through grass.
If “good” can include 50m gaps in visibility that require orienteering, excellent is really the only value which indicates a trail is adequate for the average hiker.
2 Likes
I’ve looked at that photo a few times and wondered where the path is. Never noticed the description, now it makes some sense.
Might it be possible to get consensus here on a better replacement picture that illustrates good?
To go one step further: if changes to the text are too contentious, then maybe we could work on a gallery with pictures from all over the world, with trails of different surfaces and difficulties? Like Key:smoothness/Gallery - OpenStreetMap Wiki We could limit this to examples where there is broad agreement
I honestly wouldn’t consider the faint path through grass one as "good’ in a meaningful sense of the word either. Good is essentially an intermediate value and intermediate is the beginning of bad which continues on for three more vague values.
It’s not a useful key without wording changes, but wording changes will make it (temporarily) more broken than it already is which will probably deter changes passing. I do generally agree with more photo examples being good but I’m not sure the key will really benefit from it and realistically speaking having proper photos would probably break historical usage.
I’ve soft proposed a new visibility key based on discussions and looking at different formal systems that should be more grounded in non-alpine specific reality and verifiable. Ideally this or something like it would slowly take over and then trail_visibility be used as a fallback.
I propose to change the wording in the abstract to say:
The key trail_visibility=* is used as a classification scheme for informal, back-country and wilderness paths.
That would be an update that mirrored actual usage.
Here short of 12hrs footage of trail-running over 87km distance. Here corresponding GPX of the Long Trail.
First observation: Trail visibility is not tagged much - If you are into couch mapping, this is your chance 
Second observation: In my opinion, you will end up tagging 95+ % trail_visibility=excellent, even where currently there is good or intermediate in the data. If set, good is the most used. Which clearly contradicts the documentation.
Spoiler: A 100 m in Axamer Lizum were pathless (evading snow field), I guess they are no longer 
Update on the search for actual practice: One more section of pathless terrain, videolink; Corresponding 284m OSM way was only mapped the day before the competition. As can be seen on strava heatmap, people just traverse the pasture as they please.
Interestingly, the official map has no path at all there, only in the scrub above and later up to the ridge. I know such fragmented paths from other maps as well, especially those aimed at mountaineers. This brings up an idea regarding a phrase, that has not been questioned yet in this topic:
What does map recommended/required mean? When you do not see where to continue, you can look into the map, and learn where the next path is. So you can close the gap.
Mind you, that was before the advent of routers. Routers require a continuous track. People looking at paper maps do not.
Last week I was camping and rock climbing in New York’s Adirondack Park and encountered trails with a large range of poor visibility. Most of these are in this area to the south/east of Chapel Pond and are mostly used by rock climbers to access various cliffs. They are low-traffic mostly-informal trails which do exist, but just barely.
Here’s a photo album of a few of these trails, arranged from better visibility to worse visibility.
I found that the riparian area through which these paths travel was incredibly difficult to navigate with informal herd paths, game trails crisscrossing, magnified by lush vegetation and regular water-flows erasing evidence of foot travel. For this low-lying area, visibility was the dominant challenge rather than exposure, strength, or technique.
Up on the surrounding slopes the trails often crossed large boulders and blow-downs and don’t get enough traffic to thoroughly compact a treadway. While there is a trail that is more-clear than the surrounding woods, it isn’t always easy to see here either.
I have a desire to map these approach trails as they are the only safe access to otherwise vertical terrain and they do exist, but at the same time I have a desire to indicate that they are often very hard to follow due to poor visibility.
From what I observe, trail visibility is actually lending itself much more to such paths than to what colloquially here is called hiking paths, as those are close to 100% TV=excellent, apart from some NFS Trail Class Matrix class 1 ones that are actually advertised for hiking, i.e. by a guidepost. I know such in my local area, it is a rare pleasure to get by there
@rhhs can you make sense of my proposal, to make the documentation mirror that practice in its abstract:
How would that reflect in defining values? That would cut all the ties to hiking/mountain hiking/sac_scale. It is not in the data, as far as I can tell, it certainly is not in SAC (caps) scale neither. TV is a solitary key, an OSM innovation, something some Germans came up with when trying to be finicky, it should be documented as such.
Trail_visibility has no connection to trail type or difficulty. It only describes the trail’s condition in term how clear it is. Whether it has been cleared of vegetation, well worn or even bound by some type of fencing. Tje more it is visually seperated from the surrounding terrain the better the associated value.
I agree, though there is a historical connection. @Hungerburg I agree it’s worth mentioning back-country and wilderness paths, but I think there’s nothing against tagging other hiking paths with TV to reassure the hiker that there will be no problem with visibility. Else why have the excellent and good values? There are plenty of hiking paths here in Bulgaria, even ones that are part of European long distance paths, that are trail_visibility=intermediate or worse (see the photo in my post 157 above). Of course it only really starts to matter for safety when it gets intermediate or worse.
With alpine “trails” sure, but outside of that? Not really. It’s usually extremely safe to walk through a field of over grown grass that contains a “trail” in only the most abstract meaning of the word. Same goes for trails with lots of turns or hilly areas. Same goes for over hanging branches. Maybe the trail visibility is pore because you can’t make it out beyond a couple of feet in those situations, but so what? Even a mildly fit person can duck under a branch or go around a jog in the path without either one leading to an injury. And no if someone scratches their arm on a tree branch once in a while or whatever that doesn’t make the trail unsafe.
1 Like
The “so what” is that losing the trail makes passage slow and can likely lead to getting lost. I’ve read numerous reports of wilderness rescues triggered because hikers were lost and couldn’t get themselves out of the wilderness as bad weather or darkness set in that they weren’t prepared to handle. Physical injury isn’t the only possible safety concern.
I just watched a short video yesterday in which the creator described a hiking trip where they were navigating via an app on their phone and were trying to hike out as weather turned bad. Moisture damaged their phone and then they got lost when the trail they were following became invisible in a scrubby clearing. With no paper map or GPS navigation ability and worse weather than they were prepared to handle they chose to initiate a wilderness rescue via satellite beacon rather than risk hypothermia.
Tagging trail_visibility can make it possible for renderers make it more clear which trails are going to be easy to follow and which will be more difficult to follow and allow routers to avoid suggesting routes which are likely to be difficult to follow.
5 Likes
You suppressed informal paths, that is where I see trail_visibility mapped to great extent. After all, attaching TV=excellent to a dirty shortcut on the parking lot gives the thing much more substance.
I sympathise with your endeavour, to turn this key into something, that can be more reliably determined, as you seem to be into long-distance hiking. This brings up a question: What exactly is a hiking path? Hiking is an activity, can be performed on roads too. So how to reword ma proposed abstract to mention hiking?
And yes, getting lost is not a minor cause for emergency calls.
I just watched a short video yesterday in which the creator described a hiking trip where they were navigating via an app on their phone and were trying to hike out as weather turned bad. Moisture damaged their phone and then they got lost when the trail they were following became invisible in a scrubby clearing. With no paper map or GPS navigation ability and worse weather than they were prepared to handle they chose to initiate a wilderness rescue via satellite beacon rather than risk hypothermia.
that’s so incredibly careless to rely just on a phone in an area you don’t know.
I sense hyperbole. Relying on just a phone is the normal. Heck, before I had a GPS phone, I had to relay on my senses alone!
Sure, that can happen. I’m sure there’s plenty of scary anecdotal news stories out there about any number of things that we could use to confirm whatever opinion we have about this to. Personally I’d prefer to ground conversations in something a little more substantial and relevant then edge cases that there’s zero evidence OpenStreetMap had anything to do with or would fix if only we implemented my preferred way of doing things. Maybe that’s just me though, but it’s still a fact that in 99.9% of cases no one gets seriously injured or has to be rescued because of scratching their arm on a tree branch or there being a bend in the trail even though both affect visibility. Which is what I was talking about in the message your responding to. Not wilderness hiking in the dark or bad weather. I’m sure you get the difference.
Sure, but that’s kind of a non-sequitur because any tag can make it “possible” for renderers to do all kinds of things. That’s the whole point. It doesn’t automatically that because tags make things possible trail_visibility is the best way to make it more clear which trails going to be easy to follow or not though. Let alone does it mean we even need a specific tag do to that in the first place. Sure though, tags make it possible for renders to render things. And?
And? What do you propose? Is there a better way to make it clear which trails are easy or difficult to follow? Or do you want to deprecate the tag because “we” don’t need it?
That’s my personal preference, because I don’t think it’s a very interesting thing to know about a short informal path to a parking lot. But if someone wants to add that tag, I don’t mind.
That’s up to the individual mapper to decide. I see no need to define it sharply.