To anyone who hasn’t followed all the ridiculous “tag architecture astronautics” that has been going on in this thread, it has exactly the same baggage.
3 Likes
The worldwide default access should follow the worldwide default access of the current highway=path
If it’s fine that some people classify the way as A and other as B, why we need a separation? Then obviously in the end there is no difference to the data consumers. That doesn’t mean that there are edge-cases.
The way I understand this topic: Pathway rimes with Highway (not to get an academy award, but still) – that is why the term: They are to be siblings – Pathway a new top-level key, not just a new top-level tag. So being interchangeable a bonus. Not that this will change this into anything but
2 Likes
Because there are edge cases. There always will.
Yes, there are edge cases. Same for road classification. And it’s completely fine to discuss the 1% edge cases, without that proving that the whole classification is wrong to start with.
The worldwide default access of what exactly? The new tags? Are you saying that something like highway=climbing or highway=scramble should, by default, have horse=yes and bicycle=yes? Seriously? 
One of the reasons some of us want to factor out some of these into their own tags, is exactly that the default access restrictions of highway=path are an issue.
6 Likes
Since it’s not fair indeed that I quote this in an unrelated thread, I thought I’d reply here; but Nadjita beat me to it. In a nutshell: I’m philosophically opposed to your apparent approach “since we cannot formulate hard and unambiguous criteria for classification, we should not classify at all”. It’s perfectly fine that some people classify an object as A and others and B, as long as we can agree that it’s a gray area and either class is reasonable enough.
Feel free to correct me if I misunderstood you. I maintain that it’s better that we have fuzzy, somewhat subjective classifications, for whatever purpose (and we already do for tracktype, smoothness or sac_scale) than to describe the world in endless number of, supposedly objective dimensions, which are ultimately very complex to both maintain and interpret.
7 Likes
access is regarding legal access. If in the majority of the world there are laws prohibiting horses on what you define as scramble, it makes sense to have this as default. I doubt that’s the case. If you read my previous post, I explicitly excluded climbing from highway.
I see plenty of overlapping in:
and
Especially as it’s mainly based on usage. Maybe that’s working fine in the Alps, where you have laws prohibiting bikes on most of the trails. Though that’s not the case everywhere in the world. In my experience, in the majority of the world offroad biking and offroad walking goes pretty much along with offroad horse-riding and offroad motorcycling.
Based on above definition, that should be a singletrack, because there seems to be the chance, it’s not safe for pedestrians. The same picture without mtb-rider would be perfectly describe an easy hiking path. Further above I posted a similar picture. Sorry, but I disagree with you, that that’s only a ~1% grey area.
Despite this often repeated assertion, access is not strictly limited to legality. Some mappers use it this way, but plenty of others use it to indicate practical/physical access as well. It is de facto a blend of both.
4 Likes
I seriously doubt this. It could apply to nodes (like gates), but definitely not highways. There’s nothing in the wiki to suggest otherwise: Key:access - OpenStreetMap Wiki
“Access tags pre-eminently describe legal permissions/restrictions and should follow ground truth, such as signage combined with legal regulation, rather than guesswork. They do not describe common or typical use, even if the signage is generally ignored.”
" To indicate practical (un)suitability for certain types of cycles / riders do not use mtb=* (a form of access=), but tags as smoothness= , mtb:scale=* or surface=*"
Mappers using access tags for suitability instead of the proper scale tags (like smoothness, sac_scale, mtb:scale, horse_scale) really give me a headache. Local mappers used to argue over bicycle=yes/no, since what’s ‘suitable’ varies based on skills and perception—exactly why we have scale tags.
In Thailand, trails are legally open to all vehicles by default, regardless of their difficulty. So, any unsigned trails marked as bicycle=no or horse=no will eventually be replaced with the appropriate scale tag.
1 Like
Based on our wiki definition it is.
But just assume it would be as well about suitability… climbing=*, highway=via_ferrata and highway=scramble would have a access=no since an ordinary pedestrian can’t use them. Then OSM would be kind of useless for mtb-riders, hikers,… since the router would need to ignore access completely in order to route them through their trails. Which at the same time will bring them in situations, where they are not allowed to be in.
1 Like
Yes the wiki promotes the legality only viewpoint of access tagging. The mappers who use access tags for practical/physical access as well would just say that the wiki is wrong, and it’s supposed to describe how tags are used not prescribe how they should be used.
Also note that practical/physical access is not the same as suitability. A path may be unsuitable for cycling but still possible to cycle on. This is different from a path that is physically impossible to ride a bicycle on, say because it is too narrow for the handlebars. This goes beyond mere unsuitability. The bike will not fit. If there is no law forbidding cycling on this path, the legality only viewpoint is that this path must not be tagged bicycle=no. On the other hand, the pragmatic viewpoint is that despite there being no law, it is not possible to cycle so bicycle=no fits.
I am not saying I necessarily agree with this, just that this usage exists. Physical restrictions exist in many places where no legal restriction does and in the absence of a different set of tags that are specifically not about legality, mappers use the existing access tags.
5 Likes
To express a limited width, we are have maxwidth (for legal restrictions) and maxwidth:physical (for physical restrictions). 
Yes those tags exist. And yet some mappers still prefer to just tag bicycle=no. I guess they didn’t bring their measuring tape while surveying 
. Again:
1 Like
I would disagree for scramble, beyond a reasonable level of fitness it is certainly usable by ordinary pedestrians. No special equipment is required just footwear you would use for any country walking.
Which is kind of proofing the point, that suitability should be kept out of access. I do not disagree with you. People with a certain level of fitness would be able to use it, people not having that level of fitness or confidence would not be able to use it. That’s why we are having tags to describe usability and other physical limitations.
Are mappers misusing access? Might be. Shall OSM tagging schemas promote that? Certainly not.
1 Like
I have also seen access= used for suitability in the wild. It makes sense to me, though it is against the wiki. Threfore personally I never map access for individual modes of transportation. I think there should be no “preeminently” in the wiki definition - what is the exception?
I happened today over a UIAA III scramble. Boots shown below
[Picture of dirty Oxford shoes deleted, I heard the messages]
The Swiss Alpine Club® in recent editions of its mountain hiking scale removed any mention of footwear from the descriptions. The OSM sac_scale still tells lore of mountaineering boots for UIAA I.
1 Like
No it does not. Search the OSM wiki and you’ll find this page, which is a general diacussion of climbing grading systems.
Search sac_scale and you won’t find it.
Edit: Actually you will find the text “Includes climbing pitches up to UIAA grade II” in the “SAC scale” equivalent value for the OSM value “difficult_alpine_hiking”. At no point does that OSM page say what a “UIAA I” is or what you should where on your feet.
I see it there, from T5 (verbatim) on, mountaineering boots required. In Swiss Alpine Club® scale from 2023, T5 can include UIAA II for short stretches, in OSM sac scale UIAA II only in T6 (verbatim), a.k.a. difficult_alpine_hiking.
It will be a long way to unbind OSM sac_scale from its origins.
PS: I did the UIAA III with ease in Oxford shoes and I am not a climber, I maxed out at French 5a top rope once I tried.
1 Like
I don’t care if you did it in flip-flops, but it doesn’t sound unreasonable to me for an OSM wiki page to try and document what real-world conditions an OSM tag / value combination (here sac_scale=difficult_alpine_hiking) might apply to.
I’m struggling to see if you’re actually trying to make a genuine point here or not (and if so, what it is). To recap, this part of the conversation came from this post (which in my view misinterpreted how access tags work) via this post which seems to be making an entirely reasonable point. You then … posted a picture of some shoes?
No, it doesn’t quite fit. I agree it’s an edge case, and while there’s no perfect solution, using highway=footway for narrow urban paths or smoothness=impassable on trails would be a much more appropriate way to restrict routing to pedestrians based on physical restrictions.
Then these mappers should start a discussion, reach a consensus, and update the wiki. If they don’t, the current wiki version may be used against them in conflicts with other mappers.
1 Like