not using roles on climbing relations

zbycz

And second question: I thought hard about using or not using roles on climbing relations. And in the end I think we don't have to use roles at all. All that matters is the tags of the individual members.If there is a member climbing=route/route_bottom on the crag relation - it is a crag and only these members are shown and numbered as routes. All other items (parking, etc) may be rendered outside in some "other section". It would be probably good to leave these items at the end - but I can imagine, that perhaps there would be two natural=cliffs and someone would like to sort them near the specific routes.
I added that to the wiki: https://wiki.openstreetmap.org/w/index.php?diff=prev&oldid=2892930The same is valid for area relations, but I think it is clear - I didn't add that. (edited)

9 replies

Adam Franco Tuesday at 5:31 PM

https://osmus.slack.com/archives/C04UAMU9F4K/p1756826904088359?thread_ts=1756822662.924759&cid=C04UAMU9F4K
https://osmus.slack.com/archives/C04UAMU9F4K/p1756826978370969?thread_ts=1756822662.924759&cid=C04UAMU9F4K

Adam Franco

Additionally, a named "crag" is often composed of more than one tiers of cliffs and disjoint faces, so adding the climbing* tags to one cliff isn't necessarily accurate as the level of mapping detail goes up.

From a thread in climbing | Sep 2nd | View reply

Adam Franco

I do appreciate adding the natural=cliff feature to the crag relations as a geographically-referenced holding place when there are few or no routes mapped.

From a thread in climbing | Sep 2nd | View reply

Adam Franco Tuesday at 5:37 PM

I agree that roles aren't particularly useful in crag/area relations. Roles seem most needed when there are situations where multiple items with similar tagging have a different context in the relation.For example, if there was a whitewater route relation which had two waterway=access_points, then a role on the upstream one of role=put_in and the downstream one of role=take_out might make sense, especially since the take-out for one route might be the put-in for another further downstream.Climbing crags don't really have those sort of properties and are a bit more "choose your own adventure".

zbycz Tuesday at 8:03 PM

Yes, thats the proper usecase for roles. Thanks for the reply.I will fix this in OpenClimbing tiles as well 🙂 (edited)

Stephen Tuesday at 8:20 PM

One unmet use case I've wondered about is having a way to mark the point (likely on or at the end of an approach trail path) where you enter or arrive at the crag, i.e. the landing — basically representing a target for routers. Without this information you might unintentionally find yourself at the top of the cliff instead of the bottom It would be nice for renderers to display the POI icon/label at this point instead of at the geographic center of the member nodes/ways for example

$:\+1:$ 1

Stephen Tuesday at 8:28 PM

This might be similar to how the "label" role is used in boundary relations for states/cities/etc. or we just invent a new climbing= value like climbing=base, climbing=landing, etc. if there's consensus on the need

zbycz Wednesday at 11:58 AM

basically representing a target for routers

In my thinking exactly this is climbing=route_bottom – don't you think?Or if we have the geometry of route – that would be first point of the way climbing=route. I think of climbing=route way as an oriented oneway "way". There is no need for the explicit route_bottom then.I should probably add it to wiki.

Harald Husum Wednesday at 2:10 PM

To me, having tags like climbing=base or climbing=landing is unnecessary and confusing. First of all it invites bikeshedding over exactly where to place the point. Second, as @zbycz touches on, we already can compute where the constituent routes are starting, and use that to place the label. These extra tags don't really add anything that I can see. (edited)

Adam Franco Wednesday at 4:06 PM

I think of climbing=route way as an oriented oneway "way". There is no need for the explicit route_bottom then.

I like this, the extra tags on a node in a way already tagged with climbing=route feels unnecessary.I could see the case for the node tagging on a stand alone node with no known top connected by a way though. Or maybe if a node is tagged with climbing=route then that implies that it is the bottom or at least a best estimate of the most common access point. There are situations like sea cliffs where the only reasonable approach is from the top.

$:\+1:$ 1

Stephen Wednesday at 9:18 PM

Keep in mind there are also routes that begin or end within other routes or which can only be accessed by climbing one or more other routes. There are also routes you rappel back down and ones where you cannot (or can only descend certain pitches but not others). Back in the day I tried tagging routes with oneway but eventually felt this might be misunderstood