[Tagging] The endless debate about "landcover" as a top-level tag
61sundowner at gmail.com
Fri Jun 8 22:05:12 UTC 2018
On 09/06/18 01:32, Kevin Kenny wrote:
> On Fri, Jun 8, 2018 at 4:25 AM, Mateusz Konieczny
> <matkoniecz at tutanota.com <mailto:matkoniecz at tutanota.com>> wrote:
> 8. Jun 2018 00:48 by kevin.b.kenny+osm at gmail.com
> <mailto:kevin.b.kenny+osm at gmail.com>:
> In the meantime, there is no supported tagging to show
> 'forestry' as a land use rather than asserting 'every square
> metre of this polygon is covered with trees.'
> I see no reason whatsoever to render this kind of landuse on
> general purpose map.
> The "No. You can't have that." answer, as predicted. But I strongly
> Let me take a step back.
> I'm chiefly concerned with how I ought to be tagging objects for which
> I maintain imports. I simply want to have some tagging available that
> will neither sacrifice rendering nor incur the wrath of the ontologists.
> Among these are some areas that have titles like 'State Forest'. They
> are well delineated. They are signed. While they lack developed
> facilities for recreation (typically limited to some blazed trails,
> some unpaved parking, and perhaps a notice board and register book at
> a trailhead), they offer many recreational opportunities for hikers,
> cyclists, equestrians, skiers, snowshoers, snowmobilists, canoeists,
> bird watchers, hunters, trappers and fisherfolk. They are open to the
> public, in general, whenever active harvesting is not in progress and
> the area is not newly planted for reforestation. They are at present
> tagged 'boundary=protected_area' with a protection class corresponding
> to the regulatory regime in effect for a given area.
> They are not parks, and surely not national parks, but they occupy a
> similar space in the public consciousness, because of the recreational
> opportunities they offer. According to the Wiki description,
> 'leisure=park' is clearly wrong - it envisions considerably more
> developed facilities than a patch of woodland. Still, the general
> public would surely expect that a map would include them - just as it
> includes parks.
> They are currently tagged redundantly. The tag that they all share is
> 'leisure=nature_reserve'. The term is not quite correct but it is
> nearly infinitely elastic. Since they are created to conserve land for
> sustainable forestry, they do have the conservation of nature as at
> least one objective. In my region there's a pretty broad consensus
> that it's the 'least worst' tagging that still renders.
> They are also tagged 'boundary=protected_area protect_class=6'. This
> may be inaccurate, but these areas are listed as such on the IUCN site
> as well as OSM. At present some are also incorrectly marked
> 'landuse=forest', partly because when the import was performed, the
> Wiki happened to be in a state where it described the tag as meaning
> 'land managed for forestry', and the import followed the Wiki advice.
> Nobody raised the issue on talk-us or imports when the import was
> The key aspect that makes the general public expect to see these areas
> rendered is the public recreational use. Nevertheless, that is a
> secondary use - the primary land use is that these areas are
> productive forests.
> Because many of the areas are large, they comprise ponds, mud flats,
> meadows, scrub and shrub, alder thickets - nearly the entire ecologic
> succession. Many of the ponds are cyclic and may in a few decades be
> woodland again, depending on where the beavers take up residence next.
> Even without human intervention, the land cover is not stable. Modern
> management does not attempt to extirpate the beaver (as it did a
> century ago!) but instead recognizes this as a key process in
> rebuilding the soil, preventing flooding and erosion, and supporting
> needed biodiversity. Hence, the entire area is not 'natural=wood' nor
> 'landcover=trees'. It has varied landcover that I generally map only
> for specific projects such as large-scale trail maps. Instead, I rely
> on third-party datasets based on multi-band and multi-season satellite
> imagery to identify the ecozones.
> So I return to the question: Is there 'correct' tagging for these
> areas, which are widespread in the areas that I map and are important
> to the public? What is the best strategy for keeping these areas
> rendered in the short term while still describing them correctly so
> that future rendering improvements can exploit the mapped information?
> I ask this question about once a year - and every time, a significant
> fraction of respondents give me answers that amount to, "You can't
> have that because it doesn't fit the ontology," or "You shouldn't want
> to render that because so few people are interested in that sort of
> primitive outdoor recreation", or "the fact that the land use follows
> a property line makes it parcel data, and we don't do cadastral
> information", or any number of other answers that dismiss the question
> rather than trying to answer it.
> I don't care what the 'correct' tagging is. I simply get tired of
> hearing that everything I try is 'incorrect.' or that features that
> many in the general public in my region care about are too specialized
> to render.
The best tag to me is landuse=forest ... that to me is a 1:1 fit. It is
what I use for these things that are around me.
Those who want to tag the simple presence of trees can use natural=wood
The OSM is not limited by 'the ontology'.
OSM is not restricted to someones view of what is popular. We have
provision for mapping private driveways!
OSM does cadastral information.. countries, states, counties, cities,
national parks, parks, ..
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Tagging