As someone who spent way too much time implementing topological
processing in the SFS model, I can comment on this. The "finite
point-set of intersection" rule for Polygonal geometry actually make a
lot of sense, since it makes algorithm design easier and execution more
efficient. With this rule in place algorithms can assume that an edge
of a polygonal geometry has interior on one side and exterior on the
other. This makes topological labelling easier. Also, in certain
computations you can determine an answer without having to compute the
entire topology of a geometry. (The optimized PreparedGeometry
predicates in JTS and GEOS depend on this, for instance).
Andy Anderson wrote:
Ah. Not a big deal in 2 dimensions, but I have a real problem with
this in 3D! :-) But I see in OGC 99-049 where they say:
"A Polyhedral Surface is not a MultiPolygon because it violates the
rule for MultiPolygons that the boundaries of the element Polygons
intersect only at a finite number of points."
Seems like an unnecessary restriction, but I'm sure they have their
reasons for excluding this "degenerate" case.
-- Andy
On Jun 3, 2008, at 1:58 PM, Martin Davis wrote:
Andy Anderson wrote:
I'll also note that, to be consistent, this must also mean that
MULTIPOLYGON((0 0, 5 5, 5 0, 0 0), (0 0, 0 5, 5 5, 0 0)) doesn't
have a boundary line at LINESTRING(0 0, 5 5).
Well, that Multipolygon is actually invalid - since MPs can't have
coincident line segments. So the definition *is* consistent!
I'm not sure which one that is.
It's the OGC document 99-049 (the first one you sent)
--
Martin Davis
Senior Technical Architect
Refractions Research, Inc.
(250) 383-3022
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
--
Martin Davis
Senior Technical Architect
Refractions Research, Inc.
(250) 383-3022
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
