[OSM-dev] (Multi)Polygon handling

Frederik Ramm frederik at remote.org
Fri Jul 13 18:07:49 BST 2012


Hi,

On 13.07.2012 18:43, Christian Müller wrote:
> I'd handle #2 the same as #3 - it seems weird, but it's not incorrect
> imho.  So reading along your sentence, for case #2: "[...] another,
> separate polygon, a [wood] filling out that hole.
>
> My argument is, if a mapper chose to separate them in the first place, a
> data user should assume that there's a reason for them separated, even
> if he can't deduce this reason with current OSMs data.

Before we had multipolygon relations, the segments forming the inner and 
outer ring would have been part of the same way and both tagged 
natural=wood, that's why we usually still have this exception today - if 
the inner ring is tagged the same as the outer or as the relation, then 
simply make a hole with nothing in it. If however the ring is tagged 
differently, then make a hole and fill it with something else.

> If a data user is interested in aggregating polygons with similar or
> equal tags, that's another issue..

In this "wood" example it is likely that the mapper meant to have a hole 
in the forest, not a hole filled with forest!

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"





More information about the dev mailing list