Proposal talk:Natural=rock cleanup

Place your comments here.

Landcover=bare_rock

I propose to use the key landcover=* for describing physical characteristics. --Dieterdreist (talk) 07:45, 2 August 2014 (UTC)

That has no influence on this proposal, because it suggests to not use natural=rock for simple rock surface. --Fkv (talk) 21:51, 2 August 2014 (UTC)

"remarkable"

What exactly does "remarkable" mean here, and why did you choose to make it part of the definition of a tag called simply natural=rock (rather than e.g. natural=remarkable_rock)? I'm asking because I don't want to see a repetition of the confrontation around natural=tree when mappers started to map insignificant trees, too. --Tordanik 14:24, 13 August 2014 (UTC)

The definition "significant tree" has been taken from conventional cartography, where only a limited number of trees could be mapped due to limited space in the maps and limited spare time of the cartographers. In OSM, things are quite different. We have virtually unlimited database space, and mapping freestanding trees from arial images has become easy. No need to make a preselection. We better let the applications decide on their own which trees to use and which trees to ignore. These applications can define their own criteria for significance, using various tags such as denotation=*, species=*, start_date=*, height=*, circumference=*, historic=*, and tourism=attraction.

We cannot do such a detailed tagging with rocks. Rocks are significant if the have a name and/or a special shape. This is mostly subjective. Even the definition what a rock is in the first place is much more subjective than for a tree. A tree is a a perennial plant with a single wooden stem. A rock is...what? I'd say that calling something a rock already indicates that it is significant in some way. So there's little danger of spamming the database with insignificant rocks. Sofa mappers who map every tree from arial images will hardly recognise rock formations. They only see bare_rock areas and cliffs. Mapping rocks typically requires field work.

Features like tourism=viewpoint and natural=cave_entrance are just as subjective. Even if their respective wiki pages don't say so explicitly, it is obvious that only remarkable viewpoints and caves should be mapped. Otherwise, there would be viewpoints everywhere...

Even though the word "remarkable" may be superfluous, I think that it helps distinguish natural=rock from bare_rock. --Fkv (talk) 18:09, 13 August 2014 (UTC)

That's a sensible reasoning, thank you for your answer! --Tordanik 18:48, 13 August 2014 (UTC)

Cave is not limited to remarkable ones, all may be mapped ("A cave or cavern is a natural underground space large enough for a human to enter"). Mateusz Konieczny (talk) 05:20, 14 August 2014 (UTC)

I am a caver. The cave cadastre of Lower Austria contains more than 5000 caves (with criteria: length >= 5m). If we would map all caves starting with length >= 0.5 m (that's enough for a human to enter), there would be hundreds of thousands of caves in Lower Austria. Think of limestone and granite. There are holes everywhere! --Fkv (talk) 06:56, 14 August 2014 (UTC)

Yes, Kraków-Częstochowa Upland has areas with about 6 known caves on every square kilometer (and it is expected that many small ones are undiscovered) Mateusz Konieczny (talk) 10:55, 14 August 2014 (UTC)