[Tagging] Fuzzy areas again: should we have them or not?

[stevea]

On Dec 24, 2020, at 1:55 AM, Martin Koppenhoefer <dieterdreist at gmail.com> wrote:
> I would think these are 2 different kinds of things. "incorporated municipalities" seems to be administrative entities. This is what we map as boundary=administrative, admin_level=* and it can (depending on the situation and surely varying a lot across the world) contain a lot of land that you would not consider to be part of a settlement (e.g. fields, forests, meadows, lakes, wetlands, etc.).
> Opposed to this, there is "place=city/town/village/hamlet" which is about settlements, and I would not expect a place to contain fields and forests (maybe exceptionally it can be possible, like "parks" will be part of a settlement and can contain fields and forests, parks as in leisure=park).

These are often both hard to make clear and misunderstood despite good efforts, so I offer my apologies if I did not:  incorporated municipalities, which have boundary=administrative and admin_level=*, are distinct from much smaller "settlements" (like villages and hamlets) which are not "formally organized" and are amorphous, often expressed with a place=* node.  While the former have formal boundaries and the latter usually do not, cities might contain "fields, forests, meadows, lakes, wetlands...".  The latter may, too, though as they are much smaller, it is less likely, but still possible.  Administrative areas (whether formal or not) are mutually exclusive from natural areas.  These ARE two different kinds of things, OSM creates data structures and tags accordingly.

Yes, in larger cities these "more natural areas" are often expressed as parks and nature reserves, but this isn't required.  In some cities, there are what I've called "proto-parks," a kind of "raw land" where a steep slope or other reasons why urban development is unlikely to occur there means that the city might one day develop this land into a park, but the city hasn't done that (yet) and the city keeps the land "in reserve, for future (park) development."  Here, OSM usually needs natural=* tagging (or something like it, to represent land cover), while at the same time it is inside "city limits."  These ARE "parts of a settlement," though they can either be actively used (as a park) or not (as in a proto-park).  I see no contradiction here, as tagging is mutually exclusive as I note above.