Proposal talk:Stone=*

From OpenStreetMap Wiki
Jump to navigation Jump to search



Extending/changing the proposal

Not specific enough

The key 'stone' is not specific enough to mean "an area, tagged with some other main tag, contains some stones". I see it rather as an extension to various other keys like "material=stone" "stone=basalt" to specify the type of stone material. --Mueschel (talk) 16:27, 10 May 2021 (UTC)

ZeLonewolf on Discord suggested the name of the subtag to be stones=*. Is this name more appropriate? --Gruebel (talk) 16:37, 10 May 2021 (UTC)
+1 We have Key:material and Key:surface - why not use/extend them? They also offer additional values like sand, ice etc which are not that seldom in the landscapes the proposal wants to provide a more clear tagging for, so they are more versatile. Moreover, I am pretty sure sooner or later, people will use stone=yes for walls etc. so stones will become de-facto alternative to material/surface - if we do this from the beginning, i.e. stick to material/surface, we can avoid 3 tags "for the same". --Schoschi (talk) 19:44, 10 May 2021 (UTC)

How about 'bare_rock=*

We do have the tag natural=bare_rock - that makes 'bare_rock=yes' a natural choice for areas that contain bare rock, but are not fully defined by it. --Mueschel (talk) 16:27, 10 May 2021 (UTC)

I agree. That was also one of my ideas. I just wanted to add a tag for multiple types of stone. As I replied above stones=* might be a better name. --Gruebel (talk) 16:42, 10 May 2021 (UTC)

Need unclear

Sorry, but even after reading 3 times, I do not get the core problem with current tagging - so the motivation / the need to add a new key is unclear to me. Maybe just because I'm too tired after an exhausting working day, but still, IMHO the need for the tag shall be extremely clearly worded. Within section "The current situation with grassland and stones", the values of point 3 are IMHO only to be used if no more specific value for Key:natural can be provided (e.g. bare_rock), point 2 is IMHO micro mapping, and point 1 seems quite theoretical to me - I simply choose the predominant surface, and if the landscape is around 50% stones and 50% vegetated, the new key stones=yes would also not help because both tags "stones" and "natural=grassland" are inappropriate / conflicting in my eyes. An IMHO meaningful tagging would be natural=grassland;scree (similar to how we'd tag multiple fuel types for a gasoline station, so it's clear the values exist side by side) while I am aware this is currently not allowed due to Semi-colon value separator#When NOT to use - a rule that causes creation of more and more "workaround tags". --Schoschi (talk) 19:32, 10 May 2021 (UTC)

To the second part I agree. But the idea of the subkey is to keep backward compatibility. A combination of multiple natural=* tags and landuse=* tags would be possible but difficult to implement. I had no hope that this combination had any chance of success. I think hardly anyone would agree to introduce this, because it would be "too late" or other reasons. In addition, other problems arise: in the Alps, for example, meadows often come with stones. Is a combination of landuse=* and natural=* tags useful? For when to tag: it is about areas that cannot be clearly assigned. Or in other words, areas whose appearance is significantly changed by stones. I have added many examples and the proportion does not have to be 50%. "I simply choose the predominant surface" this is exactly what I mean by point 1. This results in my problem: large areas that are tagged the same, but in reality are different. Yes point 2 could be understood as micro mapping, but on the scale of these landscapes not reasonable and thus a bad tagging style. The problem with natural=fell and natural=tundra is as mentioned that the values says nearly nothing about the landscape present. Nevertheless, they are used for such areas and thus is a bad mapping style. Now for the headline: As described, the representation of mountains can be significantly improved, because usually nothing is tagged due to the situation being so unclear. The new subtag and the changes of natural=fell, natural=tundra and the other tags ensure that these areas become clear. So people who don't want to get involved with one of the bad tagging styles can have a sensible alternative. So the main purpose is to give these areas a correct and distinct tag. I hope the amount of text does not overwhelm and can answers your questions. --Gruebel (talk) 21:47, 10 May 2021 (UTC)

Case specific solution for what is a much more generic problem

It seems to me this proposal tries to suggest a case specific solution for what is actually a much more generic problem. We classify areas by what they predominantly are - natural=grassland is an area which is predominantly covered by grass (despite the presence of scattered trees), natural=bare_rock applies to areas where the surface consists primarily of bedrock (despite some patches of soil with plant growth in between). landuse=residential applies to residential areas despite the occasional shop or office in between. What is needed for more accurate mapping and what this proposal tries to suggest for a specific manifestation of the problem is a way to indicate what secondary characterization applies to an area that is lost by the primary classification of what it is predominantly. What i suggested for this in the past is natural:secondary=*, which would mean for example

Or - in this case

--Imagico (talk) 21:16, 10 May 2021 (UTC)

I agree that this is further thought. The subkey I suggested is about areas where a new key is important. Since these places have many bad tagging styles. As a result, people are tempted rather to skip the area. With your tag I see only the problem with a combination of natural=* and landuse=* tags like landuse=meadow + natural=scree. --Gruebel (talk) 22:03, 10 May 2021 (UTC)

+1. The area should carry the tag for what most of the area is covered with. The possibly a new key of, say, "surface:secondary=bare_rock/* using the same values as the key surface=*??? Additionally the ratio of primary area to secondary area could also be tagged surface:secondary_ratio= * % where the percentage of secondary area is estimated. By not using natural:secondary and using surface:secondary it removes some values that are land forms rather than land covers and it can also be used where the surface key is used and thus add a secondary surface there too. Warin61 (talk) 09:12, 11 May 2021 (UTC)
With this proposal, I actually wanted to avoid these problems. Within one day I have now heard of at least 3 completely different ways. Each has fundamental problems. With surface:secondary I see the problem that e.g. trees are not a surface and therefore cannot be tagged as secondary landcover. That is why I decided to create a simple subkey for this specific problem. Since there are so many different ways, I had no hope that a proposal that is more comprehensive would have a chance. But maybe this discussion should be held on a larger scale. --Gruebel (talk) 10:20, 11 May 2021 (UTC)
If you view the surface with a microscope grass is not a land cover either. I think most mappers take a view from quite some distance above making trees a land cover. Where several possibilities exist a pol on the tagging list may help identify which is best. Warin61 (talk) 07:41, 12 May 2021 (UTC)