Well you’re trying to argue against proposed solutions with the status quo, but don’t really defend it. “You should just use these existing tags instead, they are totally adequate” > “This is how they’re inadequate” > “I’ll pick a clearly throwaway line and just say it’s difficult”. I’ve given some pretty clear examples where the ambiguity of SAC causes confusion, how the values of smoothness aren’t geared for hiking paths, and how surface often fails. If those are true, then it seems like some solutions would be in order (or we’re all just fine with how it is).
I’m not sure that partial information is better than none, it can lead to false assumptions. If a path is labeled as surface rock, but someone has joint issues and they don’t like all the flexing that happens on sand they’ll be disappointed when the path has sand in it. I usually don’t bother with a surface tag because I’d either have to make a way into dozens of segments or I’d just have to pick whatever seemed the most prominent.
The end of that sentence is a typo - it’s smoky here and I have a headache so when I copy and paste things around or decide to change a sentence sometimes bits and pieces get left behind. ^^
Was my take on SAC exposure more or less correct? You should be surefooted to avoid dying and to avoid bruising your nose, both aren’t great outcomes. T2 seems more like the “nose bruising” given it’s “a fall hazard” and the sample photo just shows someone walking along the shoreline of a lake. If T2 is fatal / need to rescued from a fall exposure then I’m not sure where T3-T6 go up from there heh.
I’d say the majority of trails I’ve been on recently have changed surface. Trails in needles generally start off as dirt, go into sandy washes, climb over some broken boulders, go onto a slick rock ridge, drop back down again etc. One or two short ones that lead to other trails are just dirt, but even some typical “park and hike 2 miles ones” shift from dirt to rock multiple times. Some very common dayhike trails in Yosemite go through multiple surfaces - the trail to Glen Aulin can shift from granite slab to sand to dirt every couple hundred of feet, the mist trail up to LYV goes from pavement to rocks to dirt to rocks to dirt to sand to dirt etc. These aren’t extreme examples, but ones that are pretty similar to other ones I know.
Trails below the sub-alpine in forest tend to be more uniformly dirt if in wilderness, gravel or paved if near a road, but even in the northeast you’ll often have sections of them on bedrock at low elevations (below 1000 feet). I can think of a couple casual popular trails near family in upstate NY that have this happen.
I feel like there wouldn’t be an accurate way to tag how difficult the trail is in cases those anyway. So why not just go with surface as a best guess based on what type of surface expect to be there or which ever is the hardest to hike on, like just tag the rock/dirt/sand trail as surface=rock since it’s 80% correct and leave it at that. Maybe that’s just me, but I have to assume that anyone who is hiking a granite slab trail in Yosemite probably isn’t going to have difficulty with or care about parts of the trail that are dirt. Sands probably a different scenario, but whatever. I guess you do make a valid point against using the surface tag though.
Personally, I’d link “sure_footedness” with “head for heights”, but that is just my take on the subject.
BTW: Never trust third party accounts, the summit post you linked above pretends to verbatim copying SAC documents, where they just give their own account. Where they say, “path marked to the brim” the original says “When marked, then this colour.”
The Glen Aulin trail is mostly T1 / simple walking difficulty despite going over various surfaces - the dirt is nice single track, the parts where it goes over granite slab are low angle (and the slab is grippy while being relatively even), and the sandy bits are a bit annoying but tend to be short and not impactful. There might be some areas where there are enough roots or rocks on the dirt parts to count as T2 / complex walking, but those obstacles in the surface are more impactful than what the surface actually is. That can kind of sort of be covered by smoothness, but I feel like use of that tag for non wheeled traffic is going to be all over the place.
This isn’t from that trail, but it’s just across it on the other side of the Tuolumne River. It gives a good idea of how difficult it is despite not being a dirt surface. Surface=rock isn’t always scary! (to clarify the path would be equivalent to walking horizontally near the bottom of the photo)
I have thought about just tagging trails with the most common part, but even that has issues. The Lost Canyon lop trail in Needles is mostly dirt single track, some sand when it’s near a river bank, but the middle of it jumps in both technique and surface (T1 movement up to T2/3 with some exposure, though the T3 exposure is where the movement is T2). In that case I think it’s probably worth breaking the middle part into it’s own segment as there’s a significant enough difference, but just changing it to surface=rock* doesn’t really indicate that it requires more technique, heck once on top of the ridge it’s basically T1 until you need to drop again.
It’d be nice if surface was multivalue for areas where it doesn’t really impact difficulty but still sort of tells a story: surface=rock+dirt+sand or something. I think that’s more accurate and useful vs just picking one, and more realistic and easy to parse vs creating a new way every time it changes.
I think you can need to be surefooted on unexposed terrain (say hopping along talus or boulders on a dry stream bed). I’d rate that as PG.
surefooted and with a head for heights / composure would be more R - X exposure. I just whipped up this real quick to go with my previous hiking_technique key. I had it mostly written for another project and just had to swap some phrasing around.
At some point soon I’ll try and get a us_trail_visibilty and then the three can be put together and see how effective / accurate they’d probably be,
surface=ground seems like a decent compromise, doesn’t tell you much, but tells you it isn’t pavement or gravel.
Agreed on too many options on visibility (especially when 2/3 of them are for various degrees of pathless paths, that aspect is definitely something where just copy pasting SAC to use everywhere is a bit extreme).
update: apparently the NFS has a TRAIL_SURFACE key that includes a value “NAT: Native Material Surface.” The other options are snow, asphalt, and imported compacted material. That’s pretty close to the OSM definition of ground.
But well, this is the point I am trying to make. Not that our representation of SAC T1 is wrong, but that SAC T1 is too broad for what I’d like from a hiking key-value.
Yes, to someone who does mountain hiking the examples we’ve had are about the same, but for the casual hiker there is a significant interest in knowing whether that SAC T1 is a 2 meter wide gravel way, a singletrail that atleast requires being concentrated on the ground, a forest trail where you absolutely can twist your ankle and fall if you’re not concentrated, or a way that will need either very careful navigating or good equipment in the fall because it is heavily slippery.
Absolutely. I did not mean to suggest that bringing the actual US Forest Service (USFS) classifications for their trails into OSM would be a good idea. Even if they were perfectly consistent across boundaries, they would only apply to trails on USFS land. So we’d still be left deciding how to classify the many many trails on land not owned by the US Forest Service.
I raised the USFS classification system an example from which a similar OSM system could be developed. Maybe it’s a good idea or maybe it’s not, but a simple classification system could allow mappers to record a general idea of what a trail is like when they are unable to determine the exact surface, smoothness, difficulty, visibility, etc or just don’t wish to gather that level of detail.
I was not trying to suggest that the values can be translated one to one. Different rating systems can measure slightly different things, e.g. some ratings for routes take the length of the walk into account, this obviously doesn’t make sense in OSM for a tag we put on highway objects.
But I think that comparing rating systems, thinking about what one system measures vs. another, in which respects they are similar and in which ones different, and documenting that in the Wiki, could be a useful exercise for people from outside Switzerland who are being asked to tag paths with sac_scale.
There are 730 004 entities with sac_scale in the database, of which 28 300 are in the U.S. That makes it look like the U.S.A have to catch up quite a bit.
I often don’t use it because I feel like it doesn’t really fit the path or I’d be just taking a guess (was that random way someone made T5 or T6? idk). I don’t want to create ambiguity and bad data by marking an otherwise T2 path with a few T3 sections on it as T4 because it had a mantle on it as that implies exposure and a bunch of other things (and even the T3 didn’t really map, they tended be less 10m of gain sections of broken sedimentary rock to a sandstone ridge on an otherwise flat desert trail).
I stumbled across this help post a while back while searching for SAC info and OSM. At the point where people are posting a sac_scale_rationale=* I think we can say things are broken?
Okay, thanks. I’m certainly not familiar with sac scale apart from as an OSM tag. I’ve added sac_scale to the way and those around. I’ve also added a sac_scale_rationale to explain.
This is from someone in the US, who got advice from someone in the UK who puts up notes on paths that are tagged with sac_scale because he thinks many people in the UK (even mappers) won’t be familiar with sac_scale or understand how it needs to be bent to cover use cases where it’s not a clean match. For example Way: 78591218 | OpenStreetMap has a note “hands needed on this section” presumably because it didn’t meet the other criteria of SAC T3.
sac_scale mainly refers to mountain trails while in other places of the world (Italy is an example) there are a lot of trails in flat areas or hills areas that cannot be classified properly with it.
While I think it’s useful to have some sort of standardized hiking key, I see no problem with people also tagging their local systems. They will be easily understood by locals and people familiar with the region and there is unlikely to be a 1:1 mapping with any generalized or universal key.
Where in the linked article are you seeing that? I think it just describes the SAC scale without explicitly mentioning it (yellow = T1, red and white = T2 or T3, blue and white = T4-T6). Compare the second column in the table here.
Ah yes you are right. There’s still a very large difference between the somewhat rocky single track dirt path shown on the OSM wiki for T1 and NFS Class 4-5, CAI T, etc.
Those paths being marked with SAC is interesting in its own way. 64% of the trails in Switzerland are T1, 35% are T2-T3, and less than 1% are T4-T6. In Switzerland, home of the alps etc. This matches up with trail_visibiilty etc having two values for the majority of hiking paths, and the higher four basically being for paths that have some unspecified portion of them pathless.
There may be some truth to it. I have the impression, that you overestimate T5, T6. This is not meant for the Gerbier traversal. The ropes are for glaciers in case crevasse opens, not for bolted climbing or rappel The scramble is always done unroped. If you need ropes there, this is no longer hiking.
For people, who have all knowledge about sac scale from OSM tagging of sac_scale, this might be different. Maybe the key should get renamed osm_scale?
Related more to the surface tag discussion here, though it’s from the exposure / fall_risk post.
Rock with fine gravel on it is the worst, and I’m not sure how that shows up as a surface. In the example of the Glen Aulin trail above, the worst surface is actually the slab and old rock stonework down from the bridges - covered in sand from overuse and at a moderate downhill angle.
I’ve been playing around with fleshing out YDS, though it isn’t really intended for trail/path travel, and found that thinking of terrain as loose, collapsing, skidding, or bushbashing is useful in terms of difficulty / movement. Combined with the technique / mode of travel that YDS already has and exposure/fall risk it’s a decent way to describe off-trail terrain.
I honestly don’t care much if the ground is slab, talus, forested duff, dirt etc if it doesn’t move around (I have preferences but life goes on). Modest sized low angle talus is fun to jog across, 3 C choss is just gambling against a serious injury (T4 climbing where the talus collapses down on you).
There’s some YDS Class 2 passes that feel like Class 4 - even though it’s still just “use of hands” they can have sections on this ball bearing like terrain that is fatally exposed.
King Col: Class 2 S X.
This pass should not be taken due to uncomfortably slippery terrain over fatal exposure.
The technique required is to move over steep and/or unstable ground using hands (or poles!) for balance. That’s Class 2. The fact that if you slip you die or get medivaced out instead of just falling on your ass is because it’s fatally exposed. The fact that your feet will slide when you step above that fatal exposure makes it sliding. This is why people say King Col feels like Class 4 - the proximity to the drop makes it ‘feel’ like you’re on a more technical climbing mode of travel because near vertical terrain has a lot more high cliffs than terrain 5-30 degrees in slope. 2 S X ‘feels’ like 4 because it’s exposed, not because you need to have good shoes and know how to move on rock with small holds that require some climbing techniques.
I removed the two via ferrata ratings off of hiking_technique scrambling a while back after re-reading them.
T6 includes up to UAII 2 which is an easy climbing grade, though one before protection is recommended. That is in YDS 4 terrain, which is ropes recommended. Mountaineers and climbers normally won’t rope in for it and it’s considered a scramble rating though it overlaps low YDS 5. Need is ambiguous and can depend on experience, skill, and comfort. T6 looks like it’s in a gray area of the upper reaches of scrambling, I’m sure someone has roped in on T6 terrain before.
My dad told me a story of when he was free soloing 5.8 terrain with a climber. They turned a corner and the exposure went from 400 feet to 2000 feet and the climber wanted to rope in. My dad laughed at that - it’s fatal either way, entirely psychological.
The BMC has scramble ratings that overlap climbing that it says people do free solo but recommends roping in on.
Heck, tour guides will rope in people for going up the Matterhorn, but that isn’t really climbing.
One of the T5 photos someone using a chain. Via Ferrata doesn’t really exist in the US, but if you have to use a chain is it still scrambling?
Or just use it for areas where people are familiar with it and not try to make it a global norm?
The scramble is always done unroped. If you need ropes there, this is no longer hiking.