suspicious tag combination tracktype=grade2 and surface=ground

[dieterdreist]

it seems for surface=ground, German flat land ground is assumed. In the mountains, nearly all tracks and paths are "ground", and there is nothing suspicious in the combination with tracktype=grade2, at least this is what I find everywhere in Valle D'Aosta.

There are 106000 of these in the db:
​https://taginfo.openstreetmap.org/tags/surface=ground#combinations

grade3 10,6%
grade4 9,6%
grade5 6,0%
grade2 3,9%
grade1 0,6%

[taylor.smock]

Looking at source:trunk/resources/data/validator/combinations.mapcss#L335​, this was added in r18527 to fix #22102.

• ​osmwiki:Tag:surface=ground: "It is often (widely?) understood to exclude man-made surfaces such as [surface=compacted|gravel|fine_gravel]"
  • I agree with this, since the road maintainer added something on top of the ground, otherwise surface=ground would be a meaningless tag -- it could also be applied to surface=asphalt|concrete in that case, so it would effectively be surface=yes.
  • I'd argue that surface=ground == whatever was there, unimproved. Once improvements have been made, it should be a more specific surface. In most cases, this would mean surface=ground == surface=dirt|grass|clay|sand|mud|pebblestone
• ​osmwiki:Tag:tracktype=grade2: "Mostly solid. Usually an unpaved track with surface of gravel mixed with a varying amount of sand, silt, and clay. See Gravel road on Wikipedia."

I'd like to see more clarification on what surface=ground means -- it could be tracktype=grade2 depending upon what is there (surface=gravel|pebblestone), but surface=gravel is apparently somewhat controversial for surface=ground.

With that said, I'd strongly encourage more precise tagging than surface=ground -- it isn't bad for a first pass, but what is surface=ground? If it is surface=dirt, then maybe people shouldn't be driving on it after it rains, but if it is surface=gravel|pebblestone, then it is probably OK.

Since most of the possible subvalues for surface=ground would preclude tracktype=grade2, I'm disinclined to make this change without further discussion with people who have either (a) done much more in-depth investigation or (b) have done a lot of surface and tracktype mapping or (c) uses both surface and tracktype tags in an actual application.

Klumbumbus might disagree with my stance. :)

[anonymous]

While I agree with encouraging more specific values, it should not be done with the validator, being unspecific is not an issue, it would be like asking for building=residential to put a more specific building value because it could mean anything from skyscraper to detached family home.

Generally, the usage numbers from taginfo prove that there is nothing suspicious here.

[mkoniecz]

Is this issue motivated by people using surface=ground for all unpaved surfaces including surface=compacted ?

[mkoniecz]

it would be like asking for building=residential to put a more specific building value because it could mean anything from skyscraper to detached family home

not really, as sometimes you cannot be more specific than building=residential and here you can always be (or tracktype=grade2 was wrong)

[anonymous]

I guess we are talking about a surface like this: ​https://www.settemuse.it/viaggi_italia_valdaosta/AO_cogne_valnontey_S22/foto_valnontey_S22_024.jpg
again, it cannot be suspicious because it is so widely done :)

[anonymous]

Replying to anonimo:

I guess we are talking about a surface like this: ​https://www.settemuse.it/viaggi_italia_valdaosta/AO_cogne_valnontey_S22/foto_valnontey_S22_024.jpg
again, it cannot be suspicious because it is so widely done :)

​https://overpass-turbo.eu/s/1mDb

[anonymous]

or maybe something similar to this (although this particular one could be grade3) ​http://www.rincoboys.org/GALLERY/2010/USCITE%20NON%20UFFICIALI/7)%20COLLINE%20LAVIS%20-%20VALLE%20DI%20CEMBRA/Lavis-Cembra%20(04).JPG

[Klumbumbus]

I would remove the surface=ground+tracktype=grade2 combination warning because

• surface=ground is not clear defined (wiki: "This value gives only a rough description")
• the classification of tracktype is still a bit objective (e.g. I would classify the image link above this comment as grade4)

[mkoniecz]

Still, maybe a new warning can be suitable. Is there any case at all where surface=ground + tracktype=grade2 cannot be improved?

(asking also because right now StreetComplete considers it as bad tagging and asks for surface info resurvey - if there is case where surface=ground + tracktype=grade2 is optimal then I would change it also in SC)

[mkoniecz]

BTW, thanks for notifying me about this issue!

[mkoniecz]

based on low quality photos:

​http://www.rincoboys.org/GALLERY/2010/USCITE%20NON%20UFFICIALI/7)%20COLLINE%20LAVIS%20-%20VALLE%20DI%20CEMBRA/Lavis-Cembra%20(04).JPG

Either it is not grade2 or surface=gravel or surface=compacted would be better

​https://www.settemuse.it/viaggi_italia_valdaosta/AO_cogne_valnontey_S22/foto_valnontey_S22_024.jpg

It is grade4 (or grade5 + surface=earth/=dirt)

[dieterdreist]

Replying to mkoniecz:

based on low quality photos:

​http://www.rincoboys.org/GALLERY/2010/USCITE%20NON%20UFFICIALI/7)%20COLLINE%20LAVIS%20-%20VALLE%20DI%20CEMBRA/Lavis-Cembra%20(04).JPG

Either it is not grade2 or surface=gravel or surface=compacted would be better

​https://www.settemuse.it/viaggi_italia_valdaosta/AO_cogne_valnontey_S22/foto_valnontey_S22_024.jpg

It is grade4 (or grade5 + surface=earth/=dirt)

I would never have tagged this as grade5, nor grade4, but I admit I hardly ever tag tracktypes because I do not know which is which, not even which is the criterion.

For example which tracktype is this? It is a paved surface that will be usable even after many days of rain, but you cannot use it by bike (or hardly):
​http://www.brenzonetrekking.eu/wp-content/uploads/2015/09/mulattiera-1-e1454433945405.jpg

Another surface=ground way (it is a path, but similar tracks exist): ​https://www.lucianopignataro.it/wp-content/uploads/2020/08/sentiero-delle-formichelle-escursionisti-1-800x533.jpg

Surface ground track:
​https://www.ilviaggiatore-magazine.it/wp-content/uploads/2017/03/valturnanche-piamprato-768x431.jpg

[mkoniecz]

I would never have tagged this as grade5, nor grade4

From photo it looks like soil, without gravel or anything

For example which tracktype is this? It is a paved surface that will be usable even after many days of rain, but you cannot use it by bike (or hardly):

​​http://www.brenzonetrekking.eu/wp-content/uploads/2015/09/mulattiera-1-e1454433945405.jpg

paved, therefore grade1 (and surface=unhewn_cobblestone)
this is also a perfect example why tracktype tagging is inferior to surface tagging

Another surface=ground way (it is a path, but similar tracks exist): ​​https://www.lucianopignataro.it/wp-content/uploads/2020/08/sentiero-delle-formichelle-escursionisti-1-800x533.jpg

not grade2

Surface ground track:
​​https://www.ilviaggiatore-magazine.it/wp-content/uploads/2017/03/valturnanche-piamprato-768x431.jpg

looks like surface=gravel

[taylor.smock]

Replying to mkoniecz:

BTW, thanks for notifying me about this issue!

No problem -- I pinged you specifically because you had provided the patch in #22102, and I know you work on StreetComplete.

Back on topic, reading the wiki page for ​osmwiki:Key:tracktype, it reads more like a firmness tag than anything else (as in, "how likely is a heavy object going to sink into the ground and potentially get stuck in the worst possible conditions for that road"). And maybe also "how likely is usage of this road going to affect it negatively". surface=ground would imply to me that it is more likely for something to sink into the ground in the worst possible conditions than anything else.

Maybe we should ask for clarification on ​osmwiki:Tag:surface=ground as to what is considered surface=ground? The only permissible values for tracktype=grade2 that I saw were surface=gravel|pebblestone, and I rather suspect most usages of surface=ground are closer to surface=sand|dirt|earth|clay|mud and mistagged surface=compacted than surface=gravel|pebblestone.

In conjunction with that, we might want to ask for clarification on what constitutes what grade for ​osmwiki:Key:tracktype (maybe using ​wikipedia:Penetrometer measurements -- I'll have to double check to make certain that is the right tool, I'll ask a civil engineer I know).

[mkoniecz]

surface=compacted is also a typical tracktype=grade2 value

(and in theory heavily degraded surface=asphalt surface=concrete may be)

Maybe we should ask for clarification on ​​osmwiki:Tag:surface=ground as to what is considered surface=ground?

May be a good idea, though I am not planning that (right now I flooded mailing list with discussion about drinking water features and I still need to process responses, and set up proposals)