61sundowner at gmail.com
Sun Aug 5 03:31:13 UTC 2018
On 04/08/18 22:21, Paul Allen wrote:
> On Sat, Aug 4, 2018 at 8:09 AM, Warin <61sundowner at gmail.com
> <mailto:61sundowner at gmail.com>> wrote:
> Presently landcover=clearing is fourth on the list of how to map,
> with discouraging words attached.
> That list doesn't read that way to me. It appears to conflate two
> different things: a list of increasingly-better approximations
> to mapping an area with a clearing and a list of increasingly worse
> tag use. Here's my first attempt at clarifying it...
> Increasingly-better approximations:
> 1) Ignore the whole thing.
> 2) Map the enclosing area (e.g., a wood) and ignore any inner areas
> that are not a wood.
If they landed on this page they are going to map it..
They want to map it, why else would they be here?
Don't encourage them not to map, OSM looses data that way.
> 3) Map the enclosing area (e.g., a wood) and map inner areas that are
> not a wood.
> Increasingly better tagging for mapping inner areas that are not like
> the outer area (case 3 above):
> 1) Use landuse=clearing. Very strongly discouraged, although this is
> tagging that has been frequently used
> (quite possibly because the mapper did not know of a better way of
> doing it). It doesn't render and clearing isn't
> usage of land but what covers the land.
> 2) Use landcover=clearing. Strongly discouraged, although this is
> tagging that has occasionally been used. It
> doesn't render. It is a slight improvement on landuse=clearing
> because it describes, in a negative way, what covers
> the land.
If something is not documented on the wiki . that should be fairly
strong discouragement? For example wood=clearing exists on the data base
.. but not on the wiki.
I would not mention it on a page, hopefully it will disappear in the
For the same reason I would not mention landuse=clearing nor any other
> 3) Use a multipolygon relation. The outer area (e.g., a wood) takes
> the outer role. The clearings are mapped as
> areas with no other tagging (except notes, fixmes, etc.) and take the
> inner role. This is correct way to handle clearings
> and has the benefit that it renders.
The order I have given is a compromise. I want to give the better
mapping ones before landcover=clearing so they do see them.
I want to give the simplest option first - map what is there - as that
is the easiest option. They will probably read this.
If they cannot determine what is there than I give the next option - map
the surrounds. At this point they will probably stop reading :)
If they have got this far then the most complex and best ...
Then finally the last option and least desirable .. if they have read
this far they are persistent!
If I have them in a room to speak to .. then I'd start with a multipolygon!
But in a document I try different methods, in a room I can see their
concentration and attentiveness.
> Note that it is fairly simple to convert existing landuse=clearing and
> landcover=clearing into multipolygon relationships
> however, as always, automated edits are strongly discouraged.
Yes, well. When the landuse=clearing have crossing ways, touching rings
... I have 'fixed' a few hundred of them so far .. still some to go.
Then the surrounding is confused with farmland and residential and water
.. the job of conversion is not so simple.
If it were a blank slate not having anything but landuse=clearing and
the surrounding not touching anything else... just too easy.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Tagging