On 2020-01-27 10:43, Martin Koppenhoefer wrote:
> I wonder what is the current state of admin_level on ways, in particular with
> respect to osm-carto. Historically, the recommendation was to add the lowest
> admin_level additionally to the ways that are part of admin relations (to
> help applications that render boundaries based on ways, for examples
> eliminates the need of "flattening" overlapping boundaries in case you would
> want to use non-continuous line styles).
Another related issue is the hierarchy or precedence of types of
boundaries; where an admin boundary is also the boundary of a national
park, or a political area for example. What do you put in boundary=* on
the way? I always put something in there, so my usual editor for
boundary stuff (Potlatch2!!!) shows a distinctive line, instead of a
narrow black line which is also used for millions of other types of way
like barrier=*. Admin_level is only defined for boundary=administrative,
so if the way was tagged as boundary=political then admin_level=* might
be flagged as a potential error. Hence I give boundary=administrative
the highest priority on the ways.
> However, practically this leeds to ambiguous situations, where for example
> admin_level=4 is added to islands and might be misinterpreted as
> administrative "standalone" level 4 entities (with the island name etc.).
> While a clear separation of administration and coastline could solve this, it
> would still mean continuous additional maintenance effort due to duplication
> of already present information.
I would like to take this opportunity to point out that admin boundaries
and coastline are conceptually and geographically distinct, and should
almost never coincide. Admin boundaries are typically at the low-water
mark, and sometimes miles off shore, whereas the coastline is defined as
the high-water line. I know there are different variants of "high water"
and "low water", but they are irrelevant here. The admin boundary will
coincide with the coastline where there is a vertical wall or cliff. The
island name should I guess be on the coastline; this mostly also be a
multipolygon relation for that island, so in that case the name should
be on that relation.
_______________________________________________
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging
