[Tagging] relation site <> multipolygon
lionel.giard at gmail.com
Wed Oct 3 08:40:48 UTC 2018
My main use for site relation are for historical sites to group the
historical elements of the castle or other historical site including the
wall, moat, buildings (especially the ones touching each other) and various
nodes. An example in the historical commandry of the Hospital order (in
Belgium) here :
The only historical elements are grouped in the relation (we can still see
some of the coat of arms of the order on these buildings...). I couldn't
group all the buildings into a multi-polygon like usual farms as i wanted
to get the distinction between old and new buildings (some of them have
Same thing with an historical farm
https://www.openstreetmap.org/relation/8671752 , i wanted to keep the
information about the different parts of the building (in french here
saying "old stable", "main building", "dovecoat", ...) and once again, the
different parts get different start_date.
For those two examples, i couldn't use the multipolygon due to the
buildings sharing common edge. In those case, the site relation is mostly a
"multi-polygon" whitout its limitations and the possibility to group node
when needed like the old mines where the mineshaft (or the memorial stone
indicating the shaft) are represented as a node. O
Another use of the site relation for historic object is to tag the special
heritage concerning all the site (not only one building or element).
There are also the "multipolygon" for an university dispersed in a city
like in Louvain-la-Neuve (Belgium) :
If i want to group every part of the university into only 1
amenity=university, i needed a site relation as some part are only "node"
inside a buildings (to indicate the presence of a faculty in a multi
purpose building or one of the underground parking of the university). I
tried with a multi-polygons but i couldn't add the nodes, nor the
"touching" buildings (without creating "pseudo" polygon enveloping both and
without any reality existence). Also, some buildings are multi-polygon
themselft (the "donut-like" buildings) and in at least one case, the inner
ring is not part of the university itself.
I don't find site relation difficult to interpret in those case (as it is
not for rendering !), it is usable by application like gk.historic.place
website that show the site relation objects only (i suppose they only
download the members of the relation).
Le mar. 2 oct. 2018 à 18:03, Marc Gemis <marc.gemis at gmail.com> a écrit :
> This relation combines a number of cave entrances the belong to the
> same system that is apparantly protected:
> On Tue, Oct 2, 2018 at 5:08 PM Mateusz Konieczny
> <matkoniecz at tutanota.com> wrote:
> > 2. Październik 2018 12:36 od marc_marc_irc at hotmail.com:
> > Le 02. 10. 18 à 11:46, Mateusz Konieczny a écrit :
> > > Can you link this case if that is more complicated?
> > it's a fictional example. ok not the better one.
> > take again the example you cut in the initial message:
> > a wind turbin site with a few turbines represented by a few nodes
> > I hope your solution is not to make a way for each wind turbine
> > to be able to add in into a multipolygon to describe the site.
> > it would not make much sense to make a polygon encompassing all objects
> > between the wind turbines and describe that the whole thing is a wind
> > I agree that for wind turbines multipolygon may not be feasible.
> > So far it is the only known to me case where site relation maybe is
> > (I have no experience with features like wind turbine farms so it is hard
> > for me to judge this case - that is why I skipped it).
> > _______________________________________________
> > Tagging mailing list
> > Tagging at openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/tagging
> Tagging mailing list
> Tagging at openstreetmap.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Tagging