Open andrea-perego opened 6 years ago
systems (see the original specification).
It wasn't a specification at the time, only a proposal. RFC 7946 is the only specification, not a new specification.
Both in the GeoJSON RFC 7946 and in the GeoSPARQL 1.1 spec on the geoJSON literal the fact that CRS specification is not supported in geoJSON is clearly stated. Also, in my opinion, this fact has become well-known in the community over the years.
Since the gsp:geoJSONLiteral
from GeoSPARQL 1.1 has been adopted already in GeoDCAT-AP 2.0.0, I would say that if a user wants to specify a CRS, they simply need to choose another geometry serialization like WKT or GML. I would therefore not include any additional means of specifying a CRS for geoJSON in GeoDCAT-AP 2.0.0.
I propose to resolve this issue either by
@andrea-perego what is your preference?
Perhaps OGC Features and Geometries JSON could be suggested assuming it becomes an adopted standard
@nmtoken Thank you for pointing out this initiative. Indeed, that could be also suggested, but in that case, we would also need to provide guidance on how to represent a FG-JSON literal in RDF, basically repeating what is described in #4 for GeoJSON before its adoption in GeoSPARQL - I would like to avoid that. A related discussion is actually currently being held in https://github.com/opengeospatial/ogc-geosparql/issues/249.
Until that disucssion is resolved, and a resolution included in a furture GeoSPARQL release, all we can do in GeoDCAT-AP is to point to that discussion in the note.
Resolution: No need to explicitly mention the lack of support for CRS in GeoJSON.
At the time of the release of GeoDCAT-AP v1.0 (December 2015), GeoJSON provided support to arbitrary coordinate reference systems (see the original specification).
In the new specification (RFC 7946), standardised in the framework of IETF, this is no longer the case, and the only supported coordinate reference system is CRS84 (i.e., WGS84 lon/lat) - see RFC 7946, Section 4.
Because of this, the use of GeoJSON in GeoDCAT-AP might need to be reconsidered - at least by including a caveat, and some guidance on when and how to use it.