In the brief discussion on #73 on 2025-05-26 (to be further discussed), the question came up as to how to know if a description is conforming to SRAP.
If you already have SRAP at hand, I believe there is no question: As long as 1) the profile relates shapes to classes, in conjunction with 2) those classes being used in (or entailed from) the (RDF) data, then 3) it follows that the shapes can be used to check the descriptions.
But how would one publish data and state that it (the encoded set of descriptions) conforms to SRAP, so that consumers can discover the profile?
I think the Datasets EXchange WG has proposed using dct:conformsTo relating an RDF document to a profile (or a data resource, which I presume is a closely related notion). I suspect it's wise for SRAP (and DC APs in general) to keep it simple, with one relationship (dct:conformsTo) and using the standard notion of RDF documents (which can be encoded in multiple ways); and probably not go into the more refined proposed details of XD-PROF.
Furthermore, what expectations on the published data, if any, would such a statement set? This may be a way to establish protocol and contract between the publisher and consumers; e.g. to facilitate error reporting (and perhaps even error correction on the consumer end in the meantime). This may be beyond what SRAP should prescribe, but it may be useful to make some comment about it.
In the brief discussion on #73 on 2025-05-26 (to be further discussed), the question came up as to how to know if a description is conforming to SRAP.
If you already have SRAP at hand, I believe there is no question: As long as 1) the profile relates shapes to classes, in conjunction with 2) those classes being used in (or entailed from) the (RDF) data, then 3) it follows that the shapes can be used to check the descriptions.
But how would one publish data and state that it (the encoded set of descriptions) conforms to SRAP, so that consumers can discover the profile?
I think the Datasets EXchange WG has proposed using
dct:conformsTorelating an RDF document to a profile (or a data resource, which I presume is a closely related notion). I suspect it's wise for SRAP (and DC APs in general) to keep it simple, with one relationship (dct:conformsTo) and using the standard notion of RDF documents (which can be encoded in multiple ways); and probably not go into the more refined proposed details of XD-PROF.Furthermore, what expectations on the published data, if any, would such a statement set? This may be a way to establish protocol and contract between the publisher and consumers; e.g. to facilitate error reporting (and perhaps even error correction on the consumer end in the meantime). This may be beyond what SRAP should prescribe, but it may be useful to make some comment about it.