NOTE: The galeon mailing list is no longer active. The list archives are made available for historical reasons.
I prefer URNs for advertising CRSs over listing the entire definitions (in WKT or XML), and preferably existing authorities such as EPSG can be used. In terms of URN representation, some folks from the OGC CRS WG kindly pointed me to the OGC 07-092r3_Definition_identifier_URNs_in_OGC_namespace document. In it I find a general description for not-completely-specified objects, which seems to mean (at least partially) parameterized CRSs: urn:ogc:def:objectType:authority:version:code:value1:value2 with a specific example of the Auto Orthographic CRS 42003: urn:ogc:def:crs:OGC:1.3:AUTO42003:1:-100:45 To understand this definition completely, one has to refer to the WMS 1.3 spec where they define the 3 parameters following the AUTO42003 identifier: -the conversion factor. A non-zero positive integer. See the WMS spec for more information -the lon center point -the lat center point This would seem to validate that if I want to identify a Lambert Conformal with EPSG:9802 with the origin at 25,-95 and the stdLat1 and stdLat2 at ~25 it would look something like: urn:ogc:def:method:EPSG::9802:25:-95:25:25:0:0 Note that I based the order of the parameters on the definition in the EPSG registry. One could (and should) also put in version numbers for EPSG 9802 in there as well. But as far as I can tell from how I've been reading the OGC specs, this seems like the way to identify parameterized CRSs when you are advertising a particular projection. In particular, this seems like how we would want to advertise the projection of offered coverage data available from a WCS. Does anyone have any better information on whether this is the case or if there are any potential issues? One potential issue, of course, is whether there is support for all the general projections defined in CF and beyond but I'm guessing EPSG has done a fairly rigorous job in this arena. I would also like to do research into how to properly advertise the projected (parameterized) space composed with a vertical space, such asmeters above sea level. Probably defined in ISO 19111-something. Anyone with experience doing this?
Aaron
galeon archives: