Should keys like trail_visibility & sac_scale etc exist, given difficulties with verification / some interesting thoughts on trail_visibility

This thread has turned into the renamed title above.

1 Like

Just my opinion, but I’d say no and the ones we currently use should be gotten rid of or at least extremely simplified. Whatever debate you want to have about the merits of such tagging schemes, the fact is that the consensus in OpenStreetMap leans toward not having subjective and/or unverifiable rating systems (and really in this case its an extremely glorified one) for mapped features. We don’t tag star ratings on hotels or restaurants for instance. Nor do we map the usual indicators of how safe a particular urban area might be, like crime rates or property values, all of which are less subjective then any kind of “visibility” rating (and I put it in quotes because I don’t think it actually has anything to do with visibility to begin with).

So the question is, what makes trails unique or worth making a special exception for? I can’t personally come up with an answer for that. Nor from what I’ve seen has anyone provided one. What does happen when the proponents of these types of tagging schemes are asked what makes them justified is that they will just endlessly talk in circles and miss-construe things. So I have to think there’s no actual reason or use case for the tags. At least not one outside of satisfying the feels of the person who is using them.

Like people who open notes warning other users that the person who lives down the street from them is a drug dealer. Sure, it makes the person who opened the note feel good. It’s even totally well intentioned. Does it ultimately align with consensus, the guidelines, or make the map at all better in any way what-so-ever? No, of course not. No one who is being even slightly reasonable would say it does either. Of course the person who opened the note would disagree and talk on and on for hours (maybe even for 13 years!) about why it’s a good thing and how we should just let them continue doing it. That doesn’t mean we should though. Same here. Otherwise it will be this today, ratings systems for hotels tomorrow, and completely useless “this area is crime ridden” style tagging schemes next week. Personally, I’m good on.

We don’t tag star ratings on hotels or resturants for instance.

we do

1 Like

With what tag? It’s been a while but the last I checked there was a tag for ratings that barely had any uses and it wasn’t a widely accepted practice by any means either. I’m actually pretty sure there was a discussion about it once where most people involved in the conversation said it was a bad idea that went against the on the ground rule or whatever. I don’t really care if some rando single user in Saugatuck, Michigan tags a couple of their local hotels as 3 stars or whatever. That’s not I was talking about. Nor is it the purpose of the proposal, discussion, or whatever you want to call this. You can “well, actually” me with local edge cases all you want, but the fact is there’s no “global consensus” for tagging schemes having to do with rating hotels, restaurants, or really anything else. Again, at least not that I’m aware of. Although it’s possible that’s changed between the last time I looked into it and now. I don’t think anything has though.

My leanings on this are something like “something you’d recommend sturdy footwear for” or perhaps “something you would never expect to be passable with a pram or wheelchair” should be e.g. highway=trail or highway=trek rather than path. The issue is that I don’t think something being in need of a resurface should necessarily preclude path if it is otherwise very mild. If I was in an unfamiliar city and the recommended walking path through a local park turned out to be the local mountaineering society’s test course I’d be very annoyed. Maybe my limit is wheelbarrow?

My issue is that I feel that the ‘paths’ that require three points of contact to get up and the ones that are flat paved and suitable for roller-skates are different enough to justify a different primary tag, I just don’t have a cutoff that gives a definitive answer on a brief visit, and that’s without everyone else also having to agree.

3 Likes

I probably would be to. I just don’t think that a “visibility” rating conveys that. If you want an example there’s some pretty windy downhill mountain biking trails around where I live that are tagged as trail_visibility=intermediate. I assume that’s because of the difficulty and them being windy, but I’ve ridden the trails plenty of times and I don’t think just because they have some sharp turns that the visibility is poor. So I’m left with a subjective tag that only the person who added it can get any usage out, because they are the only who know why it was added.

I guess I could come up with my own personal side scheme for trail “visibility” along with theirs, but then you’d have to agree that’s not a reasonable solution. The point is though, “visibility” can mean literally anything. Turns in a path for one person, the existence of branches over hanging the path for another. I mean, the options are literally endless. And someone looking at the map from home has no way of knowing exactly how the original mapper meant it. Compare that to something like I don’t know shoes_required=yes/no/whatever, obstacles=overhanging_plants, or potholes=yes. There’s no ambiguity there because the tags actual say what the issue with the trail is. But extremely easy “set it and forget it” tagging schemes like this one that don’t require critical thought or ground surveying to use disincentivize people from coming up with or using clearer, ground verifiable tags like those ones.

There also shouldn’t have to be multiple tagging schemes that mean essentially the same thing just because we keep going further and further up the logic chain. Otherwise you start out with something completely based in reality like potholes=yes, someone abstracts it to trail_visiblity because they think it’s easier, then another person comes along and creates a “mood” scheme so people can sort trails by how mad they are that day (I don’t know about you, but there’s certain trails I won’t hike when angry or manic because doing so would be dangerous). Then it’s essentially off to races. You could argue all of those are “useful” though. But it’s like over hanging plants —> visibility ----> mood ----> What, level of religious experience you have when you hike the trail? “Do I see god when I hike that mountain? That’s what I really want to know when I use OpenStreeMap!!” Come on.

A purely abstract analysis: what if visibility was already present in OSM, as a degree of “pathness” ?

A highway is a very visible way. A track is slightly less visible. A path even less. A non-path across a rock-field not at all (as a path) and must rely on other types of visual hints.

The underlying semantics I have in mind is that what we are concerned with is routes: where do I go? OSM’s ways with highway=* are special kinds of routes, with degrees of visibility. Other objects could represent lower degrees of visibility, even when visibility is through cairns.

2 Likes

Should we strive for a global or regional consensus for things like trail visibility & difficulty?

Yes.

People get lost trying to follow trails that aren’t easy to follow and people get hurt and die following trails that are too difficult for their skills, preparedness, or weather conditions.

Trail Visibility
As the previous thread indicated, a fine-grained deterministic 6-point scale might not be possible, but I believe that a global consensus is reachable on some large buckets:

  • The trail is so visible that it is easy to follow by novices/non-“hikers” that they can just plod along looking at their feet or chatting without wondering if they are still on “the trail” or not. Any crossing animal tracks wouldn’t confuse them because the main trail is so much more obvious.

    This is currently covered by excellent and good. A distinction between them might not be useful.

  • The trail requires periodic glances to identify where it goes. It is likely to be very obvious to anyone who knows it, has been on it before, or is an experienced hiker/mountain-biker. Novices may need to take a moment at crossing animal paths or changes in surface to determine which way the main path goes. A key distinction is that losing the path is possible for novices and there is some mental load of looking ahead for those not familiar with the path, especially in areas with many crisscrossing animal paths (e.g. grazing areas) or surfaces that don’t hold tread well (hard rock, sand, etc). Hunting for the next blaze or cairn (if available) may be needed to figure out where the path goes.

    This bucket might include trails tagged good or intermediate.

  • The trail isn’t immediately obvious or distinct from the surroundings. Even experienced hikers/mountain-bikers who don’t know the trail likely need to expend continuous watchfulness to evaluate where it goes. Crossing animal paths may be as distinct or more distinct than the path. There is a strong likelihood that novices would get lost.

    This is probably mostly falling under bad and horrible.

  • There is no visible path. A route crosses bare rock/sand/river-stone and the people are recommended to take a particular route across the area, but a particular treadway isn’t visible at all. Orienteering may be required if the non-visible segment is long.

These broad buckets focus on the effect of the conditions in terms of their likely outcome on people, which seems to be goal of most data consumers – provide an indication of applicability to various audiences and keep novices from inadvertently tackling paths that are likely to get them lost.

There will certainly be some regional variation based on trail construction practices, but “is a novice likely able to follow the path easily” shouldn’t be that subjective.

7 Likes

Is there any evidence that those things are caused by OpenStreetMap or that mapping “visibility” would at all help mitigate them? It’s fine and dandy to talk about how hikers get injured or die following trails that are two difficult, but it’s worthless as a reason to do anything in OpenStreetMap due it if there’s zero connection between the two. It’s also an extremely reductionist (I’d almost say paranoid) reason to implement a tag or map objects. We don’t map highways to save the lives drivers, we map them because it improves the project as whole if we do.

We shouldn’t be mapping highways, or whatever purely to try and save some non-exiting audiences from themselves. At least IMO doing so goes against the purpose of the project. Especially since in case as has been pointed out many times already “visibility” is subjective. If nothing else the benefits of the tag are a net neutral, in the worst case it gives the person using it a false of security that can lead to worse outcomes because everyone’s idea of “visibility” is different. I’d hate to see a story of someone dying because they thought it meant something completely different then what the user who added the tag did. Either way though, trail_visibility is really just kicking the can down the road instead of actually fixing the issue or improving anything.

This is a reference to the stars key, which is in widespread use despite concerns about verifiability. Some of the criticism of stars relates to how there are multiple scales awarded by a number of associations, so any value is inextricably tied to a particular source. Personally, I’d only tag this key based on something the hotel advertises. That’s really all a data consumer can make of this key anyways.

In my travels in Europe, I’ve noticed that many hotels and restaurants proudly display their stars on their main sign or a plaque out front. You don’t see that as much in the U.S., though hotel websites do mention their stars. (The AAA uses diamonds instead.) It’s distinct from the star rating systems that various travel booking sites also display based on user-generated feedback.

I don’t think there’s been an effort at harmonizing the various national scales into an OpenStreetMap¼ International Hotel Classification System℠, not that anything of that sort would ever get off the ground. But people are adding hotel stars based on locally appropriate schemes. One can probably surmise that a one-star hotel anywhere would be less luxurious than a five-star hotel anywhere, but beyond that it’s anyone’s guess.

At least stars is based on established standards, albeit indirectly and not in a globally coherent manner. If a trail visibility or difficulty key is based on OSM-specific criteria without reference to another standard, it would be subject to the usual OSM chaos, the tendency for mappers and local mapping communities to march to their own tune.

4 Likes

Also in most European countries star systems are based on objective factors: size of room, services available etc. I once worked with a hospital which was aiming for the “hotel” functions (catering, patients rooms etc.) to match the 3* criteria for hotels then in force in The Netherlands.

1 Like

I was mostly responding to the sac_scale / other difficulty scales part rather than the visibility part.

From a visibility standpoint I know of at least one instance where there is a definite path, then a recreation area, then the/a path continues. So far I haven’t joined them with an “invisible” path, but I’ve been tempted.

1 Like

I don’t mean to be pedantic about this, but I wouldn’t call a tag that’s mainly used in Europe and the east coast of the United States “widespread.” Sure it’s used, but it’s very Eurocentric.

That’s at least better than trail_visibility which doesn’t even seem to be coherent from one user to other, let alone follow any kind of established standards for when or how to use it. Although I do wonder what the usefulness of the star rating system is to begin with if it’s not consistent across the board since people who visit hotels probably aren’t going to be from the local, or even regional area. But that’s a conversation for a different time. I don’t want to side track this by making it about star ratings. It is a good example of the inherent issues these types of subjective, ratings based tagging schemes have though.

2 Likes

Should we strive for a global or regional consensus for things like trail visibility & difficulty?

Yes we should, and not only for trail visibility and hiking difficulty, but for all tags (that’s why we have a wiki). It is especially important for quality grading schemes such as smoothness, tracktype, mtb:scale, etc. that are somewhat subjective. Consistent mapping adds to the quality of the map, so we should strive for it, though realising that perfect consistency does not exist.

Off topic: I’m amazed how quickly discussions on this forum go off topic


9 Likes

The reason I indulged the hotel stars tangent above is that it does a good job of illustrating the tradeoffs around ensuring consistency. Because stars is a single key that’s guaranteed to be inconsistent across countries, data consumers can’t do anything more intelligent with it than display its value verbatim when selecting a search result. That’s not necessarily a bad thing: a user looking for lodging in their country will understand this system better than any other. But a user from abroad would need to adjust to this system.

I’m not experienced enough with outdoors pursuits (other than field-mapping) to comment intelligently on either trail_visibility=* or sac_scale=*, but I have often seen trails signposted with a difficulty level according to an ad-hoc system, sometimes inspired by piste difficulty markings.

If this difficulty level is useful enough to signpost, it’s probably worth exposing to the user, probably via a separate key. Otherwise, I don’t know how a data consumer could confidently derive the same succinct, intuitive rating from OSM’s SAC-esque scale. As it is, mappers are stuffing the ad-hoc ratings in name tags because they expect the ratings to appear in map labels:

Globally harmonized classifications can help data consumers make better automatic decisions: prioritizing a major city’s label over a nearby town’s based on place=*, color-coding roads based on highway=*, penalizing a smoothness=very_horrible trail, or preventing a route from going down a T4+. Meanwhile, recording verifiable, locally recognizable classifications can help users make decisions of their own: choosing a three-star hotel over a one-star motel, patronizing lgbtq=welcome establishments, or deciding against a “difficult” trail in favor of an “easy” one. It all adds to the quality of the map.

4 Likes

Probably it should prefer whatever scale is used locally, but allow conversion to others? (In the same way that OSM uses feet and inches in the US and meters in other countries.)

Here is an equivalent discussion for climbing grades:

https://wiki.openstreetmap.org/wiki/Talk:Climbing#Prefered_grading_system

In the US there’s the Yosemite Decimal System (YDS) (which is used in climbing:grade:yds_class but also applies to trails) and apparently there are different scales for Shenandoah National Park, Willamette National Forest, etc.

There’s an attempt at a conversion chart here: Hiking/Scrambling Grade Conversion Chart : General

I wholly disagree with the idea that we shouldn’t have difficulty/visibility/smoothness/solidness ratings just because they’re “subjective”. For comparison, climbing route grades are subjective, and people may disagree about the exact number, but they’re absolutely necessarily for safe climbing and planning.

1 Like

You’d have to agree that mountain climbing is a completely different thing then hiking 99% of the trails anyone is going down. It’s completely farcical to act like hiking is a super dangerous thing or on the same danger level then other outdoor activities like mountain climbing, skiing, or really anything else involving the outdoors. Especially since a good majority of deaths are caused by things that have nothing to do with the particulars of the trail and everything to do with bad planning (like not bringing enough water) or random chance (falling into a mineshaft or getting attacked by bear).

That’s completely different than mountain climbing where the dangers are inherent to activity. In no way is a “visibility” scale for paths at all necessary for safely hiking though. Let alone is one absolutely necessary. Really, it’s essentially worthless. And I say that as someone who has spent most of my life doing backwoods camping and hiking in all kinds of environments, conditions, and have tested out the trail_visibilty tag and similar ones many times to see if any of them are actually useful. They aren’t. Let alone did I ever need any of them to be safe while camping or hiking. It’s fine to support the tag, but the hyperbole and overselling about it like the tag is gods gift to hikers or something is super cringe. It also just undercuts the credibility of the whole thing in a way that it doesn’t really shouldn’t be. We can discuss the merits of the tagging scheme and improve it without acting like OpenStreetMap would litterally have blood on it’s hands without the tagging scheme.

In my region (New England USA) we have a moist climate in which trails get overgrown and become invisible due to revegetation within only a few seasons of neglect or low use. Trail visibility is a real issue as highlighted in this public radio story which documents how numerous people (generally novices at the outdoors) get lost (and some die) by stumbling onto pirate mountain bike trails and losing their bearings and not being able to consistently follow the trails. This is not to say that the danger is on the same order as mountaineering – but it is most certainly not non-existent. Not bringing enough water is distinctly a problem when what one thought would be a 2 hour loop turns into a 12-hour slog because of getting lost either due to loosing the trail or trying to find an alternative to a difficult section.

In my community many of our single-track mountain bike trails might be relatively easy to follow in mid-summer when many wheels pack the treadway in sections between rock outcroppings, but become almost impossible to follow from October through May as autumn leaves cover them and then get packed by snow uniformly with the rest of the forest floor. As someone who has regularly ridden these trails for 25 years even I sometimes loose the treadway in anything other than the height of summer.

In contrast, other trails in the same area are much more visible year round due to width and construction and are easily navigable by novices.

Just because an experienced outdoors person in one region of the world “never has trouble with trail visibility” doesn’t mean that this isn’t an issue for others and isn’t considered by them important to map. I want a map that shows the distinction between very visible trails and those that are little more than animal tracks. Having recently taken some older relatives who are novices at hiking on what I personally considered an easy hike I was surprised at how much a fear of heights and ankle twisting can be a limit to the accessibility of trails that include even mild scrambling or exposure.

7 Likes

When I’ve looked into this in the past, rock climbing is only about 30% deadlier than hiking (per participant per year).

The Centers for Disease Control has reported the third most common source of injury in the wild outdoors is hiking, second only to snow boarding and sledding. In fact, more injuries and deaths are ascribed to hiking each year than to inherently dangerous activities like rock-climbing and mountaineering.

Slips and falls make up about 50% of all hiking fatalities.

Isn’t that exactly what this tagging is intended for? Planning a day to be within your capabilities?

5 Likes

I think what your talking about is illustrated well in the first image from this post. In that case the “path” is covered by some broken branches, but it’s still clear where the “trail” is because of the clearing in the trees. If you want another example, there’s this image from Wikimedia Commons. In both cases I’d agree the “path” isn’t very clear in either one. But a “trail” is more then just the “path”, it’s the complete field of view of the person hiking, and no one hikes by looking directly at the ground all the time. What people do instead is constantly shift their focus between the foreground and background (or horizon depending on how technical you want to be about it). So sure the path in the second image is overgrown, but it’s still complete visible where to go because of the clearing in the trees at the top of the hill. Same goes for the first image, it’s pretty obvious where the path winds around the tree on the left side in the background despite the over growth. So sure, you can say path is invisible in both examples because of the overgrowth all you want, but so what? Again, no one stares directly at the ground when they are hiking.

I read through the article. This is a quote from the District Trails Manager on the Pemigewasset Ranger District from taken directly from article "And we’re walking on some trails that aren’t exactly legal. And they’re pretty clearly defined. Now maybe you can answer me this, does “pretty clearly defined” sound like the trails have a real issue with visibility or maybe the people who got lost were just disoriented because they were unfamiliar with the area? You’d have to agree that someone being lost because they don’t know the area well is different from them being lost because the trail isn’t visible.

Sure, but there should at least be some or at the least an example of where someone had trouble because of trail visibility if your going to cite that as a reason for why the tagging scheme should exist and so far all I’ve seen is a news article about clearly defined trails intermixed with a lot of hyperbole. Granted the hyperbole isn’t coming from you, but there’s definitely been a lot of it over the years this has been a topic. Yet 13 years later I have to see any solid evidence that trail visibility is even an issue. At least outside of a few cherry picked examples that probably had nothing to do with it in the first. Let alone have I seen evidence that it’s a problem that can be resolved by these tagging scheme.