The wiki page about Climbing (Climbing - OpenStreetMap Wiki) proposes to add “natural=*” to all climbing objects that are located in natural sites: routes, crags, blocks.
This makes some sense because it adds information about the nature of the climbing site. However, it also introduces confusion for data users, who apparently have difficulties in deciding which tag comes first between climbing=* and natural=*.
For instance ID chooses to depict these objects as rocks instead of crags or climbing routes. OsmAnd chooses to create two objects from one OSM node: a rock and a climbing site.
To avoid these confusions and make it easier for data users, shouldn’t we remove the suggestion to add natural=* from the wiki page?
Can you give examples of which actual tag combinations get handled by which renderers as what?
People (like me, who maintains a renderer that handle these things, but doesn’t routinely render an area you’re mapping in) can then look at similarly tagged objects near them.
Adding climbing=crag to a cliff when there’s not much other mapping happening in the area seems reasonable, but I’ve been tending to micro-map climbing areas and make relations of the routes a climbing=crag (example) rather than the cliff itself.
In some places there is a large unitary physical cliff where different sections of it have their own “crag” names, especially if large undeveloped sections between them. On the flip side, if one is micro-mapping cliff faces a “crag” may have many smaller overlapping cliffs mapped that routes traverse, especially in multi-pitch situations where it is reasonable to map intermediate cliff edges/ledges that routes pass over.
I’ve just checked, and this data user (me) isn’t confused. A regular rock (not used for climbing) looks like this, and one used for climbing looks like the icon on the right here. The one on the left is for artificial climbing walls etc. I think that OSM data is fine as is.
In Fontainebleau my problem is slightly different, but with comparable results. A crag here is a collection of (dozens to hundreds of) boulders. Consequently, picturing a crag as a single boulder can be quite confusing to end users of a map.
I note that in your example there is no natural=* on the crag, and that is more or less what I suggest to clarify things
Sorry for this, but it seems to me that if this data user does not feel confused, it may be because he is not even talking about the same thing as us
You list example of rocks, including a “climbing rock”, whereas we are not really interested in rocks per se; we are interested in “climbing spots”, which some might find moderately informative to characterize as being “made of rock” or “made of artificial materials”. The very notion of “climbing rock” does not have much significance in the data model for climbing, which is made of areas, crags and routes.
The crag in my example is not a rock at all, it is a forest area with dozens of rocks. ID pictures it wrongly as a single rock. And OsmAnd pictures it as both a crag and a rock, which is not much better.
It may be that we do not have the same understanding of “in conjunction with” in the following wiki page exercpt:
The general location of natural or man made climbing sites can be marked with sport=climbing. This can be applied to nodes for artificial climbing walls or to ways in conjunction with natural=cliff or areas for climbing halls with building=yes. It can be also applied to boulders attached to bedrock natural=rock and free standing boulders natural=stone
In any case I understand your position as being the same as mine: “natural=*” does not serve a significant purpose when added to a “sport=climbing” object, whether the wiki says so or not.
My comment about a crag with no natural=* was not a reply to one of your messages, actually. It was a reply to an example provided by Adam_Franco.
As for the example you just provided, and the various ones I found in the same area, indeed the tags are the same as the ones I initially used for crags in Fontainebleau. But as far as I can tell there is a major difference: in your example there is one boulder per crag, hence no apparent confusion (one crag, rendered as one rock). The confusion appears when a crag consists of multiple boulders as in Fontainebleau.
Maybe this will help to picture what it looks like on the ground: Le 111 - Le 111 (Gilles Cottray), which has a hand-drawn map of the crag and photographs of some of the boulders. There are 200-300 such groups of boulders in the area.
Another interesting example is here, where all the boulder routes have been mapped with one POI each in addition to the crag itself.
That understanding indeed seems to be the issue. I understand the excerpt as saying that it is possible to combine climbing tags with tags for natural objects where relevant, i.e. where the object can be mapped with natural= tags following their normal meaning, as defined in the linked wiki pages for those tags.
I don’t think the wiki ever meant to imply that climbing features must have natural tags, or that natural= tags on a climbing feature can mean something different to their standard meaning. The natural= tag should still make sense even if nobody ever climbed at that location again.
Ok, so maybe it should read something like the following?
This can be applied to: - nodes for artificial climbing walls; - ways in conjunction with natural=cliff; - areas for climbing halls with building=yes; - nodes or areas for boulders attached to bedrock (natural=rock) and free standing boulders (natural=stone); - standalone nodes when there is no one-to-one mapping between climbing features and mapped physical objects.
Note : I have taken the first step of turning the sentence into a more readable bullet list, and I am waiting for feedback before adding a new bullet point to add more flexibility.
