[geotk] WKT and TemporalCRS

[Martin Desruisseaux]

Hello Aaron

Le 07/12/10 01:05, Aaron Braeckel a écrit :
> Just to close the loop on this thread, as we discussed in person in
> Exeter it appears that I have no actual need for TemporalCRSs.  Merely
> spatial.  I do have the need for parsing GML-defined spatial CRSs, and
> possibly parameterized projected CRSs with datum assumptions
> (urn:ogc:def:method:EPSG::9802:25:-95:25:25:0:0).
The above is similar to the AUTO and AUTO2 authority CRS, but using EPSG 
operation method instead than an AUTO code. Example:

AUTO2:42003,1,-100,45  (from WMS 1.3.0 section 6.7.3.4)

Is the above EPSG::9802:25:-95:25:25:0:0 syntax something standard or a custom 
extension? In the later case, would it be acceptable to use coma instead then 
colon for separating the parameters (e.g. 
urn:ogc:def:method:EPSG::9802,25,-95,25,25,0,0)? It would be consistent with the 
existing AUTO authority and would be slightly easier to parse since it would 
separate the authority part from the parameters.


> I have been moved onto other tasks for the moment, but when I get back
> to this I think I will first look at the GML parsing capabilities
> currently in Geotk.  Parameterized CRSs make potentially dangerous
> assumptions about everything except the projected CRS itself.
For the GML parsing, I need to ask Cédric (tomorrow) what is the current state. 
In Geotk stable, there is only limited support of VerticalCRS. I think there is 
more on Geotk-pending, but I need to verify.

> Just out of curiosity, does Geotk currently have any parameterized CRS
> parsing?
There is AUTO and AUTO2 authorities 
(http://www.geotoolkit.org/apidocs/org/geotoolkit/referencing/factory/web/AutoCRSFactory.html). 
I think that it would be somewhat easy to add support for parameterized CRS 
based on EPSG operation methods. But we would need to verify if parameter order 
is determinist in the EPSG database (I think it is), and decide what to do for 
the datum.

     Regards,

         Martin