Data cube for coverage
There are multiple Use Cases for using RDF Datacube as a formalism for handling coverage data.
UC1: Describing the content of a coverage.
RDF-QB allows descriptions of the dimensions of coverage. Such descriptions allow coverage data to be discovered and interpreted, and _may_ provide sufficient metadata for certain well-known access methods to be used - e.g. WCS.
RDF-QB dimension descriptions of data support interpretation of data.
It needs to be tested whether this is sufficient for accessing data or if explicit descriptions of supported operations are necessary.
UC2: Describing slices
RDF-QB allows an abstract description of access methods and subsets of a coverage as slices through a set of dimensions
UC3: Encoding coverage data in RDF
The benefits of this need to be established. RDF-QB as a self-description mechanism would implement UC1, which otherwise doesnt imply any particular encoding requirement for the data.
Establishing a Best Practice
Is it sufficient to simply highlight the potential use of RDF-QB and provide informative examples relevant to spatio-temporal data? Are canonical abstract dimensions published as ontologies using RDF-QB (i.e. further specialisations of dimensions for space and time) necessary?