[postgis-devel] GML for empty geometries ?

[Sandro Santilli]

On Wed, Dec 14, 2011 at 10:57:47AM -0800, Paul Ramsey wrote:
> On Wed, Dec 14, 2011 at 10:47 AM, Sandro Santilli <strk at keybit.net> wrote:
> > On Wed, Dec 14, 2011 at 09:18:09AM -0800, Paul Ramsey wrote:
> >> On Wed, Dec 14, 2011 at 8:37 AM, Sandro Santilli <strk at keybit.net> wrote:
> >> >
> >> > I'm not fond of XML Schemas so don't really know what I'm talking about.
> >> > Is 'POINT EMPTY' legal WKT ?
> >>
> >> Based on 30 seconds of research I'd say go with NULL as the return for
> >> empty geometries.
> >> Yes, 'POINT EMPTY' (and all the other typological variants) is legal
> >> in WKT. (But the concept of EMPTY appears completely ignored in all
> >> discussions of WKB.) (Standards rule.)
> >
> > It strikes me that a valid entity can't be represented in all formats
> > published by the same authority.
> >
> > I filed a ticket to revert the work I did on making nice empty
> > GML geoms: http://trac.osgeo.org/postgis/ticket/1377 even if I'd
> > prefer to invent our own and bend ST_GeomFromGML into supporting
> > them :)
> 
> Yeah, we really can't do that, the GML will be validated against the
> external schemas too often for us to bend the rules.

Andrea just wrote me that this is valid (according to xml-spy):

        <gml:Point srsName="urn:ogc:def:crs:EPSG:3003">
          <gml:pos srsDimension="2"/>
        </gml:Point>

I asked him if he can find valid versions for other empties and report
to the ticket.

--strk; 

  ,------o-. 
  |   __/  |    Thank you for PostGIS-2.0 Topology !
  |  / 2.0 |    http://www.pledgebank.com/postgistopology
  `-o------'