[Tagging] Feature Proposal - RFC - Tag:shelter_type=rock_shelter
Andrew Harvey
andrew.harvey4 at gmail.com
Sat Sep 5 01:21:45 UTC 2020
On Sat, 5 Sep 2020 at 01:23, Mateusz Konieczny <matkoniecz at tutanota.com>
wrote:
> "A cave you might need a torch to explore" - note that caves may be
> smaller.
>
> In fact, some cave classifications have separate categories for caves
> small enough/open enough to be fully lit by sun and at least some consider
> rock shelters to be a type of cave.
>
According to wikipedia, "The word cave can also refer to much smaller
openings such as sea caves, rock shelters, and grottos, though strictly
speaking a cave is exogene, meaning it is deeper than its opening is wide,
and a rock shelter is endogene." which makes sense as I mentioned in the
proposal, many of these rock shelters/rock overhangs are named as caves.
Regardless I was going by the description at
https://wiki.openstreetmap.org/wiki/Tag:natural=cave_entrance and that
seems to imply it's for the exogene type not the endogene type, further an
"entrance" to the cave only really holds meaning for the exogene type.
If we can come up with a way to tag the endogene type like rock shelters,
then that free natural=cave_entrance would naturally then just be for
the endogene type where there is a large underground and enclosed cavity.
On Sat, 5 Sep 2020 at 02:19, Jmapb <jmapb at gmx.com> wrote:
> I agree that a shallow rock overhang that can be used for ad hoc shelter
> is not the same as a cave. But I'm strongly in favor of
> discouraging/deprecating shelter_type=rock_shelter.
>
I have no strong ties to that particular tag, all I want is a clear way to
distinguish these rock shelter endogene type caves from the exogene caves
which you could walk or crawl through. The shelter_type=rock_shelter tag
already has use and meets the description of these real word objects well,
so seemed the path of least resistance.
> I'm a bit strident about this because I've been personally "betrayed"
> using an OSM-derived hiking map, expecting to arrive at a shelter in
> poor weather and finding nothing. Back in civilization, I examined the
> node and discovered the shelter_type=rock_shelter subtag, but the map in
> question didn't render it any differently. Revisiting the site in fair
> weather, I found a tiny crack under a ledge that *might* have kept a
> child dry. It was very satisfying to delete that node.
>
All the ones I'm familiar with are certainly big enough to provide shelter
in poor weather. A "tiny crack under a ledge that *might* have kept a child
dry" I don't think is significant enough to be tagged as a rock shelter,
but that doesn't mean that none of these kinds of larger features which are
commonly used for shelter should be.
Obviously the map rendering can be improved, but it's against the
> general anti-troll-tagging practices to have a subtag that undoes the
> essential properties of the main tag. Because of the ambiguity as to
> what constitutes a viable rock shelter, I think
> shelter_type=rock_shelter falls into this category.
>
I don't see how this subtag undoes any essential properties of the main
tag, after all amenity=shelter is described on the wiki as "A small place
to protect against bad weather conditions", which a rock shelter is
exactly. Nothing on the amenity=shelter wiki page says that it's only for
man-made features.
>
> I'd suggest natural=rock_shelter as a replacement tag.
>
I'm not tied to any one particular tag, but just looking to further
formalise a tag that already has some documentation on the wiki and already
has extensive use in OSM.
If that tag was used, unless the wiki definitions are changed for
amenity=shelter to be strictly man-made it would not be wrong for someone
to tag amenity=shelter + natural=rock_shelter, which I guess is fine, as
one is mapping the actual natural feature and the other is saying how it
can be/is used.
On Sat, 5 Sep 2020 at 06:21, Tom Pfeifer <t.pfeifer at computer.org> wrote:
> +1 for going into the natural key
> My expectations to amenity=shelter would be something purpose-built,
> which is true for all subtypes except that rock_shelter
>
Nothing on amenity=shelter wiki page says that it must be purpose-built and
not naturally occuring.
On Sat, 5 Sep 2020 at 08:47, Martin Koppenhoefer <dieterdreist at gmail.com>
wrote:
> amenity=shelter is not about being purpose-built but about something being
> used for a purpose.
> I would not use the natural key with a value like "shelter" or
> "rock_shelter", as this is about the purpose, a specific use of the
> situation, and not a description of the physical situation.
> From the point of view of shelter tagging, shelter_type=rock_shelter seems
> a valid approach and I would go for it.
>
> You could still double tag it with natural=cliff_overhang (or whatever
> describes the feature) if you like.
>
Agreed.
On Sat, 5 Sep 2020 at 09:25, Mateusz Konieczny via Tagging <
tagging at openstreetmap.org> wrote:
> amenity=shelter is not about being purpose-built but about something being
> used for a purpose.
>
> +1 - I am 100% fine with amenity=shelter with rocks acting as a shelter
> (as long as usefulness
> is as good as purpose made shelter, not like with this children-sized tiny
> overhang that was mentioned)
>
Agreed.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20200905/51758954/attachment.htm>
More information about the Tagging
mailing list