[postgis-devel] SRID / Spheroid / Units

[Mark Cave-Ayland]

Paul Ramsey wrote:

> I am assuming that the GEOGRAPHY type will include an SRID and that if 
> one is not defined, WGS84 parameters will be assumed. Unlike GEOMETRY, 
> any measuring operation (length, distance, area) against a GEOGRAPHY is 
> going to need to know the spheroid parameters in order to return an 
> answer, which implies a spatial_ref_sys lookup.
> 
> I am not fond of the idea of having to do a lookup, because it's going 
> to have to happen so often. Particularly for some well-known values, I'd 
> like to short-circuit that process. Unfortunately that cuts against the 
> grain of the spatial_ref_sys concept, which is that the srid numbers are 
> in fact internal to the database, and not universal (srid 4326 could be 
> anything, it's not guaranteed to be WGS84). Any advice or guidance? 
> Maintain the pristine theoretical underpinnings of spatial_ref_sys, or 
> bow to the universal use we have promoted, which is to expect the values 
> in spatial_ref_sys.sql (and hence, EPSG numberings) to prevail.
> 
> P.

I really don't think this will be an issue, as it's exactly the same 
principle as the PROJ.4 cache - so only one lookup per SRID is required 
per query. So I'd say go ahead with allowing a SRID to be specified for 
the geography type. I suspect the hardest part will be parsing the 
sphreoid text into some kind of internal structure...


ATB,

Mark.

-- 
Mark Cave-Ayland - Senior Technical Architect
PostgreSQL - PostGIS
Sirius Corporation plc - control through freedom
http://www.siriusit.co.uk
t: +44 870 608 0063