[postgis-devel] Motion: vote for considering SRID <= 0 as "unknown"

[Sandro Santilli]

On Tue, Oct 04, 2011 at 10:11:59AM -0700, Chris Hodgson wrote:
> Sandro Santilli wrote:
> >On Tue, Oct 04, 2011 at 03:54:28PM +0100, Mark Cave-Ayland wrote:
> >>My only comment would be what do the various standards (SQL-MM/OGC)
> >>say about SRID values? As long as they don't specify anything, I
> >>guess we have a bit more freedom to interpret the specification as
> >>fits best with PostGIS.
> >
> >All I can find about SRID in OGC SFS 1.1:
> >
> > Before a geometry can be constructed and inserted into a table,
> > the corresponding row for its SRID must exist in the
> > SPATIAL_REFERENCE_SYSTEMS table, else construction of the geometry
> > will fail.
> >
> >There's no mention of the possibility for a geometry to have _no_ srid.
> >The SRID field in spatial_ref_sys and geometry_columns is of type
> >"integer" (but there's no case of "unsigned" in the whole document so
> >that isn't necessarely a sign of the fact that we should support <0).
> >
> >The SQL/MM doesn't seem any different in that reguard. In both
> >there's no concept of "unknown" srid.
>
> I thought the entire motivation for switching to the unknown=0
> semantic was for some kind of standards compliance... if this isn't
> in any standard, why are we doing it again? (I'm seriously asking
> the question - what is the positive side of this change?)

I think is conformance to some "de-marketing" standard, but could I
bring your attention to the motion being discussed here ?

It's not about returning 0 or -1 or null from ST_Srid but rather for
considering the whole set of SRIDs <= 0 as semantically equivalent and
thus convertible into a single value (to be discussed which) with meaning
"unknown".

What do you think about that ?

--strk;

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html