Maybe not as often as for other highway types, but then why do we need tracktype? tracktype is supposed to help map users take decissions on whether to use a track or not, isn’t it?
1 Like
Believe it or not, a lot of human transport still happens on foot :-).
2 Likes
Someone asked me to use the same method to visualize the correlation between surface and smoothness, so I’ll share the result here, as it may be useful to others:
| Surface | excellent | good | intermediate | bad | very_bad | horrible | very_horrible | impassable |
|---|---|---|---|---|---|---|---|---|
| paved | 71715 | 18527 | 16849 | 6353 | 1201 | 483 | 143 | 192 |
| asphalt | 1276256 | 1648678 | 357862 | 60408 | 6117 | 1488 | 381 | 272 |
| paving_stones | 163478 | 399081 | 123754 | 8694 | 788 | 210 | 70 | 339 |
| concrete | 106562 | 122229 | 72663 | 10283 | 1824 | 452 | 92 | 391 |
| chipseal | 57 | 699 | 329 | 92 | 2 | 3 | 0 | 0 |
| concrete:plates | 4284 | 8738 | 8572 | 3620 | 448 | 93 | 32 | 35 |
| metal | 291 | 759 | 461 | 155 | 24 | 22 | 6 | 59 |
| concrete:lanes | 366 | 1402 | 1349 | 1092 | 383 | 65 | 8 | 13 |
| sett | 2151 | 21148 | 59450 | 26950 | 3178 | 84 | 34 | 32 |
| wood | 790 | 4349 | 3787 | 1463 | 549 | 421 | 279 | 531 |
| fine_gravel | 411 | 9497 | 50245 | 24145 | 3179 | 477 | 101 | 107 |
| grass_paver | 374 | 594 | 2141 | 2063 | 177 | 64 | 15 | 21 |
| compacted | 920 | 35175 | 91938 | 62571 | 9628 | 2343 | 655 | 397 |
| cobblestone | 258 | 939 | 2063 | 5369 | 1186 | 250 | 92 | 43 |
| gravel | 502 | 8421 | 48809 | 91912 | 23533 | 8199 | 2400 | 1069 |
| unhewn_cobblestone | 48 | 134 | 654 | 1888 | 1136 | 114 | 53 | 41 |
| pebblestone | 55 | 595 | 2203 | 5167 | 3304 | 513 | 317 | 157 |
| sand | 50 | 5113 | 5843 | 13980 | 10307 | 5506 | 1427 | 496 |
| unpaved | 251 | 5826 | 29663 | 90887 | 109747 | 35920 | 7958 | 4553 |
| dirt | 255 | 1541 | 13344 | 44455 | 43536 | 17376 | 9027 | 3010 |
| grass | 119 | 1255 | 10584 | 33942 | 25942 | 13604 | 8337 | 2772 |
| ground | 524 | 7594 | 47649 | 152027 | 142313 | 67554 | 35360 | 12560 |
| rock | 5 | 393 | 183 | 519 | 534 | 569 | 475 | 476 |
| earth | 10 | 113 | 770 | 3567 | 3601 | 3741 | 2081 | 628 |
| mud | 5 | 63 | 240 | 1358 | 1380 | 1794 | 1921 | 875 |
The data was extracted from the same planet.osm file from 2 February 2026.
4 Likes
I am wondering if you are checking the tagged data against their respective images, or just the images alone.
For the first pair, it seems like the depth of the potholes is different. The first image might have potholes that are too deep even for a normal car with normal clearance (18 cm as defined in the wiki), so maybe it represents a situation that would be tagged smoothness=horrible instead of smoothness=very_bad.
The second pair looks very similar. When assessing its suitability for a city bike or a wheelchair, I would imagine these modes by law would use the sides of the road, not the middle, so I think the second image is a better description of smoothness=intermediate. The wiki does not contain that nuance explicitly, it only says that the vehicle should be able to traverse the route without significant risk of damage. In the first image, these modes would have to dodge some small potholes, which puts them at another risk (of accidents).
As for the last image, has it been used to describe smoothness=bad? Sett pavement can be bad (unsuitable for city bikes or wheelchair) if the gaps are both wide and deep, or if it’s got bumps from lack of repair, like this (which is actually quite common in my area, could go take a photo and replace this with a copyright-free image):
Something that looks a bit more even like this is still somewhat risky for those modes:
But both are relatively fine for normal cars, although a bit uncomfortable.
Personally, I think the biggest issue with sett is reduced grip when wet. This is a property of the material (stone) and shape (flat), but not something that can be described with smoothness or tracktype.
Oh, so we should have smoothness=bicycle and smoothness=car :-D???
The current vaguely-named tag values were formalized in terms of vehicle usage and approved in 2008. In July 2009, someone added value-like names in parentheses in the “Usable by” column of the smoothness tag article indirectly via a template, perhaps trying to inspire mappers to adopt these more accurate values instead of the vaguer ones still used today. That never really caught on. But in essence, based on wiki descriptions that have been unchanged in the last 17 years, that’s basically what values like smoothness=intermediate and smoothness=bad mean. The wiki says that:
smoothness=intermediateis usable by “city bike, sport cars, wheel chair, cruiser/touring motorcycles, Scooter and all below”smoothness=badis usable by “trekking bike, normal cars, commuter/standard motorcycles, Rickshaw and all below“
Exactly as in the approved proposal. There’s probably less than 1% of smoothness left from a time before the approval of this proposal, here’s its chronology:
1 Like
My original requirement was for a tag that would loosely describe the tracks surface. I ride/walk/drive on mud/sand/grass, depending where I am. The explicitly stated materials and density were too detailed to tag off, so rough categorisation was a away of making things a little bit easier to implement and scale, given the status of the wiki map at the time which required a lot of work to even get main roads completed.
Secondly, for the UK specifically there was a challenge, which was that highways (rights of way) such as a foot’way’ could be on a track or a path. Having the track type used alone allowed for a cycle’way’ or foot ‘way’ to in essence exist on a separate layer, allowing for a physical map and an access rights map to be separate but rendered together clearly. It’s not unique to the UK, and having travelled considerably in this time since, I can say that the same need for differentiation was universal although in different forms.
What was intended however by me and the few editors that spoke about it at the time, doesn’t mean anything ‘should’ or ‘shouldn’t’ be happening now in it’s use, as things have presumably evolved. The 'visual scar’ on the land is one way to grade it, the roughness another, and the degrading being another. Originally, given how basic everything was, it was very much all of those in one. ‘how good it was to travel along’…simple as that.
It is really good, in this thread, to see an effort to make different systems align with rules to compliment. Originally when I posted the request, it had 5 pictures on it. 1 for each. It roughly aligns still with how it is today.
Edit: And here is that picture from that wikipage around 20 years ago:
10 Likes
And that picture is pretty much how I’m still tagging it today.
I wonder if reverting to that 20-YO version of the wiki would be best way forward 
5 Likes
It at least needs a second set of pictures for rocky dry climates. I am always at a loss in say Turkey (or even Spain).
2 Likes
As a matter of fact, the Wiki page has not substantially changed in last 10 years, including the image, and I think it shouldn’t. (Most apparent changes are actually whitespace and footers).
That much I agree; however, it’s unlikely that we’ll reach a consensus about which pictures and what they should represent, so we’ll have to have another dozen threads, including at least four polls. 
1 Like
While I’m saddened that you totally misinterpreted my genuine confusion, I thank you for at least attempting to assume good faith, even if end result leaves something to be desired 
. I want to assure you that there are no angry mastodons here.
Nothing to say. Thanks for trying so hard to make no sense of what I wrote
I actually tried hard to make sense of what you wrote, but it turned out to be impossible without being privy to additional information that you provided below[1] – thus the post expressing my confusion and asking for clarifications.
Thank you, that other thread would indeed be an absolutely necessary prerequisite for understanding most of that message of yours I replied to (and with that reply, seemingly have inadvertently annoyed you with). I’d ask you to try reading your post in this thread without having any knowledge of what you wrote in that other thread, and I’m pretty sure you’d too see why it would be very confusing.
In any case, I hope you’ll too come to conclusion that no offense was intended by me.
Personally, I think I’ll unsubscribe from this thread, as tensions running that high can’t possibly be healthy; and even if they somehow magically resulted in perfect tracktype solution (which seems rather unlikely to me at this point anyway), it still wouldn’t be worth it IMHO.
OSM (as any other hobby) should primarily be fun, eh? This ain’t doing it for me. HAND.
which I thank you for! It would be sufficient to provide just that link for background info to help me make sense of your words in this thread ↩︎
1 Like
It may also be that the page hasn’t changed much simply because some fairly basic questions about the current wording remain unresolved. For example:
- the introduction mentions “how well-maintained” a track is, but there are no concrete maintenance criteria described
- the relationship between firmness and maintenance is not explained if maintenance is implied
- it’s unclear whether firmness refers to the surface layer or the overall load-bearing structure
- the illustration shows features such as the middle strip as a differentiating element, yet this is not discussed in the text
- at the same time, people recommend tagging concrete lane tracks as grade1 regardless of the middle strip, because only the wheel contact surfaces matter
If the documentation cannot clearly answer these kinds of questions or acknowledge regional variation, then leaving the wiki unchanged will continue to create confusion for both mappers and data consumers.
2 Likes
Well, in some other topic I stated, that I consider such and such a mandate to do just that and got one thumb down.
My personal favourite: Poll (hopefully better): Tracktype never dies! - #54 by SomeoneElse – I’d phrase it like that:
Key tracktype provides a simple classification of how convenient a track is to travel along in a vehicle. See table below for description of what the values each are.
PS: Regarding confusion, I guess Poll (hopefully better): Tracktype never dies! - #68 by ezekielf answers a lot about that. At least now I understand why people confused tracktype with a sibling of smoothness, that only maps a single scientifically provable (I doubt) aspect, but got confused when reading the linked documentation, which does not say such.
1 Like
What I suggested was even simpler - I didn’t have the “in a vehicle” part in it!
The reason wasn’t because I wanted to be particularly inclusive to non-vehicular traffic - it was more about keeping the volume of text down on the highway=track page to make it easier to read.
You imagine trackype also on paths? Sometimes I wish there was something like that…
There is! Roughly 1% of highway=path have a tracktype, and slightly more than 1% the other way around.
Something like this? Flip the numbers around (5-1 
1-5), replace “class” with “grade”, and it starts to look familiar.
Edit: Here is a compressed upload of the linked file above (since usda.gov links appear to be not working in number of countries)
USFS Trail Class Photo Examples.pdf (3.0 MB)
Some of us walk on tracks. And very much care about tracktypes. Grade1 and 2 tend to be unpleasant. Possible 3 as well. 4 and 5 are great.
2 Likes
