Add constraints on compound CRSs

[Piero Campalani]

There are known constraints when compounding CRSs, in the geospatial context.

SECORE should inhibit the composition of incompatible CRSs, and in the first place of identical CRSs:

​http://kahlua.eecs.jacobs-university.de:8080/def/crs-compound?

1=​http://kahlua.eecs.jacobs-university.de:8080/def/crs/EPSG/0/4327&
2=​http://kahlua.eecs.jacobs-university.de:8080/def/crs/EPSG/0/4327

The possible compounding include:

• Geographic 2D + Vertical
• Geographic 2D + Engineering 1D (near vertical)
• Projected + Vertical
• Projected + Engineering 1D (near vertical)
• Engineering (horizontal 2D) + Vertical
• Engineering (1D linear) + Vertical

Source: ​http://www.ihsenergy.com/epsg/guid7_1.pdf

[Dimitar Misev]

Yes I know about the definitions (also in ISO 19111), but isn't that a bit limiting? It gives max 3D compound CRS.

[Piero Campalani]

Yes it is limiting, but I meant to check validity of CCRS with regards to the geospatial ones:, which should be:

• gml:ProjectedCRS
• gml:GeodeticCRS
• gml:VerticalCRS
• gml:EngineeringCRS
• gml:GeographicCRS
• gml:GeocentricCRS

Then e.g. gml:TemporalCRS can be appended anywhere.

[Bang Pham Huu]

This ticket covers this ticket ​http://www.rasdaman.org/ticket/150, so I will do this ticket as soon as I understand it correctly. My question is, I've seen one can create 5D coverages with compound from 4 "CRS" (GeodeticCRS + TemporalCRS + VerticalCRS + Vertical CRS), is it valid with regard to ISO 19111? or are duplicated compound CRS (2 Vertical CRS) are valid?

http://localhost:9090/def/crs-compound?
1=http://localhost:9090/def/crs/EPSG/0/4326
&2=http://localhost:9090/def/crs/OGC/0/AnsiDate
&3=http://localhost:9090/def/crs/OGC/0/Index1D?axis-label="elev"
&4=http://localhost:9090/def/crs/OGC/0/Index1D?axis-label="sensor";

[Dimitar Misev]

Let's skip this one for the moment, it definitely needs to be discussed first.

[Dimitar Misev]

I think #150 is clear and can be done first.

[Dimitar Misev]

@Bang: this is the ticket you mixed up with #150. Note my earlier comment:

Let's skip this one for the moment, it definitely needs to be discussed first.

[Bang Pham Huu]

@Dimitar:

I've known this ticket before I did in ticket 150 and I think this ticket should be discussed clearly (as you pointed out).

Beside my earlier comments (#5), There is a new type of gml:CompoundCRS ​http://www.rasdaman.org/ticket/679 and this makes me confuse about the combination of this CRS type instead of concatenating from single CRS as ticket description.

[Dimitar Misev]

Regarding #670 - that is not a new type, it's the same old compound CRS (same definition structure), but it has an official CRS code. In petascope we most often generate a compound CRS on the fly with crs-compound, which then doesn't have an official code. #670 is about enabling petascope to support the officially defined compound CRSs, those that do not have crs-compound in the URL.

[Peter Baumann]

this is an interesting field for geoinformatics, but no agreed solution available on which we can build. I tried that initially and modestly in OGC and was pointed out that such a thing does not fly. There is no known set of rules. Also, we have much more pertinent tasks.