[postgis-devel] Re: Q3C

Martin Davis mbdavis at refractions.net
Thu Mar 19 10:46:58 PDT 2009


Generally I would agree, Mark.  But AFAIK there's some serious R&D 
involved in providing geometric functions on the spheroid (e.g. 
relationships, overlay functions, buffer).  I'm guessing this would be a 
3-4 month effort just to add this to JTS, let alone port it to GEOS.  Or 
do you have deeper insight into this problem?

I have no clue what MS or Oracle do to provide geodetic functions.  
Actually that's not quite true - in this document

http://download.oracle.com/docs/html/A88805_01/sdo_cs_c.htm

 in sec 5.7.2 it says that buffer, centroid and convexhull are supported 
by preforming them in an "implicitly generated local-tangent-plane 
Cartesian coordinate system".  Sound familiar?  I suppose this indicates 
that they do overlays in geodetic space - which still seems like a big 
chunk of rocket science.

Mark Cave-Ayland wrote:
> Paul Ramsey wrote:
>
>> My current thought is an automagical planar projection into an
>> appropriate UTM zone or polar stereographic system, based on the
>> centroid of the feature. With the extra complication of first looking
>> into the projection cache to see what the last projection used was,
>> and if it's close enough, using that. Hopefully that way, in the case
>> of st_function(geog1, geog2) the two arguments would be both cast into
>> the same plane. This would magically allow the geography module to
>> make use of all the geometry functions, and only break for really
>> large objects that exceed our chosen planar areas.
>>
>> P.
>
> I must say that I'm not a particular fan of this approach, since I 
> can't see a way for the user to know what level of accuracy they are 
> getting for any particular calculation. Plus having some operations 
> randomly fail because they cross certain points/areas is just going to 
> be a nightmare to support.
>
> I'd like to see a geography type implemented as a pure angular 
> calculation set with its own set of related functions derived from 
> angular arithmetic, plus with its own geodetic indexing scheme. I 
> don't see the value in just re-hashing what we have with a different 
> type so we can claim geodetic support, when in fact the reality is 
> very different.
>
>
> ATB,
>
> Mark.
>

-- 
Martin Davis
Senior Technical Architect
Refractions Research, Inc.
(250) 383-3022




More information about the postgis-devel mailing list