I would probably say “follow the trail”, but English words aren’t OSM tags. I just think that the OSM concept of highway=path is a linear feature that a person can follow. In other words, it’s conceptually the same if you’re following a dirt rut or a sequence of cairns.
I think, what you describe not a feature of tag highway=path but a feature of key highway=*
Why? What is fundamentally different to the hiker whether they are following a dirt rut or some rock piles?
Rock piles (cairns) generally created on sites with great visibility, not necessarily where one would go by. Just giving an indication of where to aim.
We are pedestrian, not on rails.
Quoting myself from above:
I haven’t been up that one, but from what you’re describing (“there’s clearly an obvious linear route”), and without getting into the other discussion about paths vs scrambles, surely there’s something worthy of a highway tag there.
1 Like
That’s a fair point as long as the mapped paths are better than the alternative. Sometimes that seems to be a judgment call, but us mappers seem to have more inhibitions about making those judgment calls than do professional cartographers in their armchairs.
OSM has quite a bad reputation among some land managers in the U.S., because word has gotten out that the parade of hikers getting into trouble or destroying delicate habitats are relying on this thing called “© OpenStreetMap contributors”. We’re working hard to improve this situation, but you know, it’s a big country. Quite in contrast to the boasting in this thread, the reality is that most of the trails in this part of the world have been mapped remotely from very comfortable armchairs, leading to some less than ideal outcomes no matter what our documentation says.
Sometimes mappers trust maps too much. This highway=path was originally traced from a decades-old USGS topo map[1] and presumably some corroborating hints in satellite imagery. The trail fell into disrepair and was abandoned. This sign was intentionally removed, but someone made a cairn to replace it.
Our coverage of this trail seems to have enabled the AllTrails user community to get away with confabulating an attractive “optional unsigned shortcut” for some time. It became an almost weekly site of helicopter rescues before the media picked up on it and both sites were corrected. Sure, more detailed tagging of physical characteristics would’ve made us feel better about ourselves, but there was a lack of groundtruth in the first place.
I get a pit in my stomach thinking of all the TIGER-imported abandoned mining roads that I saw looping around the mountains of southwestern West Virginia some eight years ago, that in a panic I retagged from highway=residential to highway=track or highway=path. I had no idea if these paths had been repurposed for the state’s up-and-coming mountain biking and ATV scene or if they had been completely abandoned decades ago. With only crude satellite imagery full of shadows, I couldn’t even tell if these old roads were even in the right place or actually traipsing up and down the mountainside with abandon.
But hey, at least an OSM-powered self-driving car wouldn’t attempt to traverse it anymore.
The “USGS Topographic Maps” layer available in iD and JOSM is a curated selection of USGS quadrangle maps from various pre-digital series. The layer doesn’t include newer series because the USGS started from scratch when going digital, ditching their famed but decrepit road coverage in favor of TeleNav (now TomTom) and all but eliminating trail coverage. ↩︎
2 Likes
To me, this is the key point in the discussion. This question remained open in that other thread about the paths I added. I had the same impression as you. A path on a map shows you the way even if there is not much on the ground you can follow. You know that, by following that path on the map, you can reach the end point.
But this is not the understanding of many mappers or map users out there, sadly (or fortunately?).
So, this is what we are trying to work out here. One of the proposals was to have a separate tag for those cases where this is not clear.
This is also a very valid view when viewed from the opposite perspective (and I find it correct, too, at the same time). Strictly speaking, most people will not see a path on that photo. Path, as a physical feature, is not there. But there is “something” - a trail, route, way, alpine path, off-path, pathless…
1 Like
Thank you for this post. That finally made me understand your position.
One thing is not clear - do you see value in distinguishing paths that are right of ways/best ways to summit but not really visible on the ground from paths that a re visible?
This is a very nice example of a path=trail or something like that and definitely is not pathless (though it might be seasonally variable). When you talk about a hypothetical but common example of a path that is not discernible on the ground, should it be rendered the same? If not, is there another way to indicate that it should be rendered? I think I have seen something like designation=public_right_of_way. Would not UK centered users use such tagging? This line of thinking feels very UK centric to me - are there other countries where such old unmaintained rights of ways are a thing and somebody cares about them?
This definitely should be marked on a hiking map (I think it should not be mapped on Google Maps or somethign similar that is not topographic, if they actually used OSM data). But still I would like the information that no physical trail is there. Or do you think such information is not beneficial?
(from there I would like to remind that the current model largely fails in making this distinction, especially in general purpose maps)
Ease of orientation. I might be ok walking on the dirt path in a mist but rock piles are much trickier. under such conditions.
In that case, I think you were right and the mapper that removed the path from the map was wrong. Doubly so if you personally visited the site and mapped it and an armchair mapper removed it!
In my area, the topography is the result of glacial retreat from the last ice age. It is common to have sections of bare rock in otherwise forested areas. So a trail will follow a dirt rut for awhile but then transit over bare rock for awhile, and then resume the dirt rut. If it is a longer stretch of bare rock, it is common to paint trail blazes (markings) on the rock. We always map the highway=path across these.
One of the more prominent places where this occurs is on mountain tops which, in New England, tend to be above the tree line only for the tallest mountains in the area.
For example, see Mount Mansfield, which is Vermont’s highest peak, and also a place I’ve personally hiked. Above the tree line, the path is mapped straight over the bare rock and hikers follow cairns and blazes. It’s a very popular trail and hikers have no trouble following it. The famous Appalachian Trail and Long Trail go over this mountain.
If we were to remove highway=path from this section and replace it with some other tag that is not immediately understood by data consumers, it would make the user experience for OSM and OSM-based maps immediately worse for any application that consumes map data on Mount Mansfield.
Now, on the flip side, awhile back @Hungerburg proposed highway=scramble to tag scrambles, which is in-between what we commonly consider a path and what would qualify for tagging as a climbing pitch. There was a fair amount of support for this, but ultimately not enough to make the change.
In the case of scramble at least, if highway=path started disappearing from data consumers in locations where there’s a scramble, that might not be a wrong outcome. Scrambles are far more of an outlier on the path spectrum than are ones that have low visibility.
4 Likes
I think this is a misunderstanding:
When a path is officially marked (and so visible), it would stay path. So =pathless is meant for paths that you wold not know are there if you did not have a map. Now you will probbaly say these should not be mapped at all, however there are a lot of them in OSM and there is big reluctance to remove them. I have so far not noticed a reaction to this from you, unless I missed it.
Yes, if people want to do so, and it could certainly be useful in some circumstances. But I think that’s secondary to whether or not there is a usable route that can be and is used by those on foot - which is what I’d associate highway=path with. As I’ve already said, I think the local visibility aspect can be captured perfectly well with trail_visibility=. And if you’re concerned about difficulty of use, there’s also sac_scale=, surface=* to record additional attributes.
Should it be rendered: yes, if it can be and is regularly used by people. Should it be rendered the same: that depends on the purpose of the map being produced and how many different path styles you think your users can cope with. If you had three styles of path, then maybe: “paved”, “unpaved but clearly visible”, and “not clearly visible” would be three categories to use. But some people may prefer to merge the last two and instead use sac_scale to prioritise difficulty over visibility. If you had only two styles, then you’d need to chose which to lump together. The choices would depend a lot on who your map’s audience is. For general use, I’d say paved/unpaved is probably the most important distinction (which is what OSM Carto does). For maps aimed at mountaineers, you might make a difference decision.
designation=* is about legal rights, not visibility or physical accessibility. You can have e.g. designation=public_footpath on a route that is paved, unpaved, not clear on the ground but obvious from signs/end-points, not really clear at all on the ground, overgrown to the point of being unusable, or completely blocked by buildings, ditches or streams. The last couple of these you definitely wouldn’t want recorded as a usable pedestrian route, so I wouldn’t tag them with any physical highway=* value. While designation=* is used by many UK mappers, the intricacies of Public Rights of Way law is probably not understood by a most of the population, and probably not by the majority of UK mappers either.
2 Likes
Not certain about that. I searched the web for Mount Mansfield and found a gallery, this picture here looks to me like sac_scale=alpine_hiking: IMG_3840 | Hyrum Wright | Flickr and a nice easy scramble too 
Curiously, the markers are the same colour I have seen on photos of SAC maintained advanced routes.
Oh, I have hiked that exact location on Mansfield. It is not really much more difficult than climbing a steep flight of stairs (although it is equivalent to many flights of stairs). I’d rate it T2 or T3 and I’d call this a hike and not a scramble. There is no meaning to the marker colors, in the US they are used arbitrarily (blue trail, red trail, orange trail, etc.). I’d say it’s very difficult to assess sac_scale from photos alone.
This is really where the scramble discussion went off the rails - what is an easy hike for one person, is a hard one for others, and I think people struggled with finding the cut-off.
This kind of situation is incredibly common in Hawaii hiking. At elevation, it is nearly always raining and the mud is incredibly slippery. You have to use the sketchy rope to go up and down safely:
I think this kind of thing is more different from ordinary paths than ones with low trail visibility. It’s closer to highway=steps! Except there’s no steps. Really, you could call it stepless steps. How about highway=stepless + surface=mud?
I just got reminded on this Von der Station Eigergletscher | Berg- und Alpinwandern | Schweizer Alpen-Club SAC – graded by the gods themselves. In openstreetmap the route is even graded one grade higher, and I think the picture shows the key spot.
PS: This one blazed by SWW (not the SAC).
Dunno, looks like an ordinary path to me, possibly T3 or T2 (but I agree it is hard to judge on the basis of photo alone, if it is steep it could be T4 too).
Pictures can be deceiving! You would not be able to remain upright unless you were wearing crampons and using the rope. Otherwise you will faceplant on the mud and slide down. Walking on Hawaii mud is as slippery as ice and it has to be experienced to be believed. Note, using crampons in this way WILL violate the warranty, but all the local sporting goods shops carry them. And this is on an island that has never seen snow. It easily qualifies for T5 when experienced in reality.
Yet another reason why in-person survey is the gold standard for mapping a hiking trail.
5 Likes
I fullly concur that in person survey is essential. I would not be tagging based on pictures :-). However, SAC scale allows for technical aids - their presence lowers the grade (the same thing without a rope would be graded higher than with a rope). Is this trail pictured in the muddy condition? It looks dry to me. I can imagine the tempting not to walk on the trail but on the vegetation next to it further driving erosion must be strong. (essential feature of T5 is that if you fall, you might die, I think - I am not however opining if that applies to this trail).
The rain has a tendency to pockmark the surface, especially if the trail is less traveled. So pictures can be deceptive. The heavily trafficked ones show the telltale signs of people slipping and sliding up and down. Also, these are generally atop knife edge volcanic ridges - more or less straight down on both sides. Deaths from falls on Hawaii trails are unfortunately common.
2 Likes
Yeah, I would think that scrambles are a subset of pathless. Then there could be just rock, sand, grass, etc. so it includes any non-visible path.
I’m still struggling to figure out what are the criteria for subdividing paths. On that paper map, it is the width. On French maps, there is “the hard part”.
I actually agree with @ZeLonewolf on this one—I might be worth to revisit the scrambling/mountaineering angle.
From a developing country perspective, like Thailand, the ‘pathless’ idea just doesn’t fit. Outside a few managed national parks, trail conditions change constantly.
There are usually no signs, cairns, or markers; trail segments often become ‘pathless’ simply due to low usage, overgrown vegetation, mud, or sand. Trails can practically disappear for weeks, months, or even years.
Some involve navigating through dry streams or rivers, and after the next flash flood, any traces might vanish, leaving obstacles that completely alter the route.
But the more popular hiking, MTB, or enduro routes will still use these paths, and they’ll likely be tagged as highway=path. So, a ‘pathless’ or ‘don’t map it’ approach doesn’t make much sense when conditions are so fluid.
Another idea that come to my mind would be to distinguish between maintained and unmaintained/informal trails—basically, “if it’s not official or maintained by a national park or hiking group, use it at your own risk”.
3 Likes