[Tagging] Avoid using place=locality - find more specific tags instead

Joseph Eisenberg joseph.eisenberg at gmail.com
Wed Apr 17 13:30:38 UTC 2019

Re [1] Grande Cariçaie
Looking at the relation, all I see is “type=group” and name=“Grande
Cariçaie”. If you load the members, you see that each is a way,
fortunately the names include “Reserve Naturalle” so that helps.

But how am I to know why this relation is? Its not a nature reserve or
protected area by itself? Where does the name come from, if it isn’t
an official protected area with some sort of shared administration?

If it were a type=multipolygon with leisure=nature_reserve, or a
boundary with protected_area, it would be clear what feature this name
refers to.

[2] Group of sculptures
I’ve seen sculpture gardens, which can be mapped as an area.
These sculptures are all in one row, so they could almost be mapped as
a linear way instead of as separate nodes.
But I agree that a relation type for a group of nodes could be useful
for a number of things, including art installations that are scattered
over an area and can’t be perfectly represented by a node, line or
However, I would still like this to work like other relations and
ways: the tags need to be on the object (the relation). In this case,
when I open the relation all I see is a list of nodes, with no tags. I
have to select one of the nodes to find out that it's a sculpture,
then check all the others to see if they are they same.

On Wed, Apr 17, 2019 at 4:24 PM Markus <selfishseahorse at gmail.com> wrote:
> On Tue, 16 Apr 2019 at 16:26, Joseph Eisenberg
> <joseph.eisenberg at gmail.com> wrote:
> >
> > > @MarKus: Regarding the tagging of islands or lake groups (clusters), I've
> > > already begun to use the type=group tag and hope that someone will push
> > > OSM-Carto to render such relations in the future.
> >
> > It will be very difficult to handle such relations in osm2pgsql, the
> > tool that is used to import the database for rendering, as long as the
> > group relation can include other relations, ways, and nodes in one
> > object.
> >
> > Is there any reason that lake groups cannot be tagged as multipolygon
> > relations? These are already handled by most database users, including
> > Openstreetmap-Carto.
> It's not just about groups of lakes. There are other groups, where the
> individual elements either have no name or individual names, for
> example this group of natural reserves [1] or this group of sculptures
> [2]. The group of sculptures consists of nodes, thus a multi-polygon
> relation doesn't work for it. And for other groups that could be
> mapped as multi-polygons (such as the group of natural reserves), this
> would mean that we would need new tags for about every existing tag
> (at least for about every tag that is used on areas).
> [1]: https://www.openstreetmap.org/relation/8856988
> [2]: https://www.openstreetmap.org/relation/8961321
> _______________________________________________
> Tagging mailing list
> Tagging at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging

More information about the Tagging mailing list