[Tagging] The endless debate about "landcover" as a top-level tag (was: Re: British term for municipal greenery?)

Kevin Kenny kevin.b.kenny+osm at gmail.com
Wed Jun 6 15:10:44 UTC 2018


On Wed, Jun 6, 2018 at 10:03 AM, Mateusz Konieczny <matkoniecz at tutanota.com>
wrote:

> landuse=forest is used for tree-covered areas,
>
> not for a forestry landuse.
>

Which leaves us with no reasonable way to express the latter. That's why
this discussion keeps coming up - at least around here, there are
significant areas that have forestry land USE, that the locals want to see
on a map (and yes, I'm willing to render my own!) and may or may not be
coterminuous with the tree cover.

There is a significant number of vocal people in previous rounds of this
discussion, which I concede has been beaten to death, who do no even
recognize that it is reasonable to want to map these area. Often they deny
that such things are observable in the filed. When I've pointed out that
some of these areas are signed, I've even had private mail saying that the
only way to approach that is to map each individual sign and leave it at
that because signs don't delineate an area feature.

It's certainly possible to render the cross-cutting concerns of land use
and land cover.
https://kbk.is-a-geek.net/catskills/test4.html?la=42.5731&lo=-74.1489&z=14
shows one possible example (Sorry about the clutter of point features and
the imperfectly distinguished colours - I'm not 100% satisfied with the
rendering, but it's good enough for my use at the moment.  The landcover is
at fairly low resolution from an external source; the land use is from OSM.
The sorting out of the land use is rather complicated, since it's kind of
deduced from 'leisure=*', 'natural=*', 'landuse=*', and
'boundary=protected_area' tags; it's fairly nasty and not altogether
reliable.

So we have available to us:

landcover=trees - seldom used, but available and unambiguous
natural=wood - controversial, what qualifies a woodland as being 'natural?'
There's next to no land anywhere on the planet that has not been managed by
humans in some way.
landuse=forest - asserted to be synonymous with landcover=trees, but has a
natural-language meaning that it designates a forestry land use.
boundary=protected_area - I suppose, but sorting out 'this is protected for
forestry' requires parsing not only protect_class but also the
natural-language 'protection_title' or other nonstandard tagging.
leisure=nature_reserve - At least this one renders, and a lot of things are
nature reserves. Including working forests, maybe, I suppose, but this
smells of tagging for the renderer.

Nothing really fits "This land is used for production of forest products"
These go together, in the cases that I've mapped, with the secondary uses
of game management and public recreation, but I'm happy to use access tags
to describe when the public may enter and whether they may bicycle, ski,
snowmobile, hunt, trap, or fish.

It's not just the state forests, either. I haven't tried yet, because of
the tagging problems, but I'd like to map the conservation easements (where
the landowner agrees to confine his use of the land to forestry, in
exchange for tax relief) and related land uses, because they are
significant to some things that I use maps for, but there's no agreed-on
tagging. I concede that this desire strays close to mapping cadastral data
(on which there appears to be near-universal agreement that OSM is the
wrong repositlory), but it has fewer privacy implications, being a matter
of public record, and is considerably less voluminous than a full cadastral
map. There is precedent here, as well: when we map a park, a nature
reserve, the grounds of a school or hospital, we are mapping cadastre!

Anyway, I'm frustrated, because the people who make the assertions that
"Tag X is not for use Y" too often do not offer a better alternative, and
when challenged on the fact, some even simply fall into saying, "that sort
of thing doesn't belong in OSM." I would rather think that if something can
be seen in the field, has a well defined spatial location or extent,
mappers want to map it, and renderers, routers, navigation systems,
statistical analyses, and so on want to use it, it's fair game.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20180606/113cb3e3/attachment-0001.html>


More information about the Tagging mailing list