Abstract Test Suite GeoJSON mentions SRS's that cannot be specified

[redmer]

The abstract test suite says in § A.3.4.2:

The IRI \http://www.opengis.net/def/crs/OGC/1.3/CRS84\ shall be assumed as the SRS for geo:geoJSONLiteral instances that do not specify an explicit SRS IRI.

Yet an explicit SRS IRI cannot be specified in a GeoJSON literal, per Req 26, which says:

RDFS Literals of type geo:geoJSONLiteral do not contain a SRS definition. All literals of this type shall, according to the GeoJSON specification, be encoded only in, and be assumed to use, the WGS84 geodetic longitude-latitude spatial reference system (http://www.opengis.net/def/crs/OGC/1.3/CRS84).

I had hoped that this could refer to JSON-FG but alas that does not extend the "geometry": {} JSON object, but rather specifies a new co-ordinate "place": {} object.

[situx]

Hi @redmer thanks for picking this up. To clarify:

• Yes GeoJSON literals may only have one CRS, namely CRS84 as per the definition of GeoJSON as a format.
• And yes, I think the wording in the abstract test suite is very poor there, as you cannot define another SRS I think we will correct that shortly. In fact it might be a copy/paste mistake from other literals which indeed contain an SRS URI. I will create a pull request to change it.

As for JSON-FG: That would require a JSON-FG literal which we did not define in GeoSPARQL 1.1. If there is demand for that we could think about adding it for next revisions.

[situx]

We changed the wording for the better, so that issue can be closed now