GeoDCAT-AP [PR]: Recommendations against the use of dct:conformsTo to reference systems (spatial and temporal)

One of the open issues of GeoDCAT-AP is the mapping with spatial and temporal reference systems. None of the RDF vocabularies under consideration provided suitable candidates. As the SDW WG is working in these issues, the GeoDCAT-AP WG decided to provisionally use the property dct:conformsTo for this purpose. Afterwards, the GeoDCAT-AP WG planned to update the GeoDCAT-AP accordingly with the recommendations that the SDW WG will make.

 

Some inconveniences have been pointed out on the use of dct:conformsTo since it is also used for mapping other information such as conformance to a standard or implementing rule. Therefore, there is a risk on overloading the term. Feedback was received against the use of the property dct:conformsTo to reference spatial and temporal systems:

Proposed solution. The use of a more specific property was suggested by John Blower and Simon Cox. It is planned to make a proposal on how to model this information, taking into account the GeoDCAT-AP requirements and submit it to relevant standardisation groups, such as the SDW WG and the W3C Locations and Addresses Community Group, where similar issues have been discussed.

 

The issue is identifying what is strictly necessary, in terms of properties and/or classes, to specify a reference system, and to query metadata records in order to get this information. If possible, this should take into account that de jure or de facto standards may be released in the near future, and therefore there's the risk of adopting a solution that cannot be easily adjusted to ensure compliance with such standards / best practices.

1. Keep using dct:conformsTo, but "type" the resource corresponding to the reference system. This could be done by using one of the following properties:
   1. (a) dct:type [3], with a code list value defining the notion of (spatial / temporal) reference system. Furthermore, the INSPIRE Registry operates a glossary including these notions. E.g.:
      a:Dataset dct:conformsTo <http://www.opengis.net/def/crs/EPSG/0/27700> .
      <http://www.opengis.net/def/crs/EPSG/0/27700> dct:type <http://inspire.ec.europa.eu/glossary/SpatialReferenceSystem>.
   2. (b) rdf:type [4], with a class defining the notion of (spatial / temporal) reference system. This class may need to be defined, since it is not available in the reference RDF vocabularies taken into account. E.g.:
      a:Dataset dct:conformsTo <http://www.opengis.net/def/crs/EPSG/0/27700> .<http://www.opengis.net/def/crs/EPSG/0/27700> rdf:type :SpatialReferenceSystem .
2. Using a specific property to denote (spatial / temporal) reference systems. This property may need to be defined (possibly as a sub-property of dct:conformsTo), since it is not available in the reference RDF vocabularies taken into account. E.g.:
   a:Dataset :spatialReferenceSystem <http://www.opengis.net/def/crs/EPSG/0/27700> .

• New terms don't need to be defined and a reference code list (the INSPIRE Glossary) could be used.
• If specific properties and classes will be recommended / defined in the future, and adopted at the EU and/or international level, it would be easier to adjust the GeoDCAT-AP specification and existing implementations.

 

Vote

Option 1(a) is the option that has received more votes among proposals so far.

See the mailing thread [1] and [2] to get more detail.