[Talk-GB] Group relation problems and alternatives (from Tagging canal staircase locks and others)

[Tom Crocker]

Based on the responses to using the group relation for a group of locks I
wanted to understand more about these views and get suggestions for a
specific problem.

What I would like to map 'better' are named woods. There's a local wood
called Spring Wood [1] with distinct areas of needle-leaved and
broad-leaved trees. I think ideally I could map these different patches and
the overall named area. It gets more complicated because that wood and a
bunch of others are collectively referred to as Esholt Woods [2]. I've
currently used a cluster relation* for Esholt Woods and ignored the
different patches of tree-types.

One approach would be two layers of cluster (or group) relations but I get
the general anti-relation sentiment when a simple area will work.

Another approach would be to have areas for the named woods containing the
smaller patches of wood. These could also be tagged as natural=wood, but
this seems obviously wrong. They could be tagged place=locality, except
they clearly relate to an extant feature and the wiki asks us to avoid
this. I'm not aware of other tags I could use.

Nodes for the names would be possible but it seems a shame to not indicate
the area and I still don't know what the appropriate tagging would be.

More creative ideas include creating a multipolygon thin 'ring' of wood for
the named woods, surround the other woods (don't worry, I'm not going to do
this).

The benefit of a relation seemed to be that it reduced redundancy and said
what it was on the basis of its members, but I can see that as soon as you
have different member-types interpretation would be very difficult. Also,
it's not currently supported by anything so doesn't really help.

Any suggestions?

Thanks for reading. Happy mapping!

Tom


[1] https://www.openstreetmap.org/way/90709491
[2] https://www.openstreetmap.org/relation/12148252

* I know, I suggested a group relation for the staircase locks, I thought
that was what I'd used here and can't remember why I chose a cluster
relationship. As far as I can tell it's the same idea as a group relation.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-gb/attachments/20210824/647ab033/attachment.htm>