[Tagging] The endless debate about "landcover" as a top-level tag

Warin 61sundowner at gmail.com
Wed Jun 13 22:42:28 UTC 2018

On 13/06/18 23:01, Kevin Kenny wrote:
> On Wed, Jun 13, 2018 at 8:00 AM Warin <61sundowner at gmail.com 
> <mailto:61sundowner at gmail.com>> wrote:
>     On 13/06/18 19:48, Martin Koppenhoefer wrote:
>>     2018-06-13 11:44 GMT+02:00 Mateusz Konieczny
>>     <matkoniecz at tutanota.com <mailto:matkoniecz at tutanota.com>>:
>>         13. Jun 2018 11:42 by dieterdreist at gmail.com
>>         <mailto:dieterdreist at gmail.com>:
>>             2018-06-13 11:36 GMT+02:00 Mateusz Konieczny
>>             <matkoniecz at tutanota.com <mailto:matkoniecz at tutanota.com>>:
>>                 Obviously - ownership would be recorded in owner tag
>>                 (rarely done for obvious reasons) and
>>             what are the obvious reasons not to record if land is
>>             owned by the public or privately owned?
>>         Complicated and boring to survey, limited usefulness of this
>>         information.
>>     the usefulness of knowing the land ownership depends on the
>>     jurisdiction.
>     The amount of time and effort in obtaining the information may be
>     beyond the mappers tolerance.
>     The may want to map other things with that time that they see as
>     much more important and usefull to them.
>     If the name ends with "State Forest" you know who operates it, it
>     is ultimately the State Government .
>     Access to state owned forestry areas is normally public.
>     Closed or at least restricted when logging or there is a special
>     event on - like a car rally.
>     These things are generally understood, infrequent and so not
>     normally mapped. These exceptions are what OSM does not cater for
>     well.
> Yes.
> I don't do a lot of landcover mapping, because I render my own maps 
> with third-party landcover data. I do landcover for detail mapping in 
> my own neighbourhood, for producing large-scale trail maps of specific 
> small areas, or to override trouble spots in the third-party 
> database.willing o
> For landuse mapping, what chiefly concerns me is recreational 
> opportunities.
> To this end, I maintain a few imports of public lands in New York 
> State, as well as mapping various public-access lands that are in 
> private hands. I do try to map access in places where it's 
> complicated. Some of these lands are managed for forestry - and I have 
> no tag available to indicate this. Neither 'natural=wood' nor 
> 'landuse=forest' appear to mean anything more than 'shade this area 
> green on the map, and draw trees on it.'
> If this discussion reaches some sort of rough consensus, I'm certainly 
> willing to do mechanical edits updating the few thousand areas that I 
> imported. (Mechanical edits to correct systematic errors in imported 
> data are, as I understand it, acceptable.) I'm not very happy with the 
> use of those imports as evidence of 'this is tagging practice' - the 
> current import is more 'least worst' tagging that will remain 
> consistent with current rendering and with imports in neighbouring 
> states. I'm NOT willing to retag with mechanical edits if the price of 
> the retagging will be that the State Forests, Wildlife Management 
> Areas, Watershed Recreation Areas, and so on will disappear from the 
> main map.
> What I want:
> Showing that land is treed ('natural=wood' or the proposed 
> 'landcover=trees') is easy enough. 'landuse=forest' appears to be 
> synonymous with both 'natural=wood' and 'landcover=trees' and so isn't 
> useful to me, although I've tried consistently to use it to indicate 
> designated land use and not landcover. The result has been rendering 
> gaffes where trees are overlaid on water - but they don't bother me 
> excessively, since most of those ponds will have trees again in a few 
> decades, as human and beaver remodeling of the land shifts elsewhere.
> 'landuse=forest' to designate landcover is unworkable. As Warin said 
> in an earlier post, a piece of land has one use. (I oversimplify; land 
> may have secondary uses, for example, land managed for forestry with 
> public recreation as a secondary objective, but NEITHER of those 
> implies that a particular square metre is or is not tree-covered.) An 
> object like https://www.openstreetmap.org/relation/175474 is an 
> example. The land USE is, correctly, military - the land is used, for 
> instance, for live fire exercises. The land COVER for large parts of 
> the Academy is trees - what an ecologist would call 'temperate mixed 
> forest'.
> That treed land cover is contiguous with Bear Mountain/Harriman State 
> Park (which should be boundary=national_park, but that's a different 
> argument), Black Rock Forest (private land open to public outside 
> certain seasons), Storm King State Park (managed, effectively, as 
> leisure=nature_reserve), Storm King Art Center (amenity=museum - an 
> outdoor sculpture gallery in a partly-wooded setting), and various 
> private holdings (where I'm not trying to tag land use).
> So, what I'm after is: some tag that I can use for something like the 
> International Paper tract in the Adirondacks (not mapped because (a) I 
> haven't got to it, and (b) the tagging would be just too 
> controversial). It is owned by a private common-stock corporation. It 
> is managed to grow trees for paper (as you might imagine). It is 
> ordinarily open to the public to hike, ski, snowshoe, and so on except 
> in areas where active logging or reforestation is in progress. Several 
> public trails traverse it. It is not a nature reserve of any sort. It 
> is private forest land. (I oversimplify; it's within the boundaries of 
> the Adirondack Park, and so regulated tightly by the Adirondack Park 
> Agency, which is a public-private partnership. It's complicated, but 
> to a first approximation, it's 'private land open to the publlic.')
> So I'd like to see landuse=forestry foot=yes ski=yes 
> operator='International Paper' etc. for this parcel. The really 
> important aspect is the access, but it's a little strange to have 
> access tagging without the object's representing some sort of tangible 
> thing.
> I'd earlier have said, 'landuse=forest', but that tag has become too 
> ambiguous. There's no good tag available to me. I'm not convinced by 
> the arguments that we can't introduce a new tag: "It will just be 
> abused the same way that landuse=forest was" is simply a counsel of 
> despair. "It conflicts with existing practice." Any new tag conflicts 
> with existing practice. "There are too many areas out there to change 
> them." I'm volunteering to change the few thousand that I entered, as 
> needed. I'm sure I'm not the only one who's willing to put in the work!

I'd think there is around 1,000 here that I can change. About 600 in 
'my' state and the rest I hope to find mapped in OSM other states.

At the moment for rendering green with trees there is natural=wood. And 
that can be combined with landuse=forestry.
To me the natural=wood tag is a 'tagging for the render' thing, and I 
don't like it. But it may be a work around until landuse=forestry is 
accepted and rendered in a different way to 'there be trees'.
I'd not add the tag landcover=trees, in the long term I'd remove 
natural=wood from landuse=forestry to keep landuse=forestry clean of 
'there be trees'.

There is also landuse=logging which I suspect is really forestry ... 
over 47,000 uses of that tag.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20180614/9ff1ea2d/attachment.html>

More information about the Tagging mailing list