Open hoehrmann opened 8 months ago
the profiles vocabulary [1] allows links to resources with role qualifiers. treating this as a polymorphic equivalent or subclass of Distribution solves this problem. Im not aware of anything else that does it in a canonical fashion.
see https://www.w3.org/TR/dx-prof/#related-dcat
In DCAT-AP, dct:conforms-to is noted to express such information. There is no specific guidance to how a dct:Standard should look like and what it should contain as it really depends on the publisher of the Dataset or Data Service.
Note that the dct:conforms-to in DCAT-AP Dataservice is connected to the protocol, the interaction, e.g. it is SOAP,REST JSON, WFS, etc. While the structure of the data is documented in the associated Dataset.
dcterms:conformsTo
is for »An established standard to which the described resource conforms.« I might say a CSV file conforms to RFC 4180 or the SDMX-CSV standard, but this is about human-readable documentation, dcterms:conformsTo
is not suitable for that.
@hoehrmann, personally I think this is more on how strict you interpret the notion Standard. If a city creates a specification for its datasets, then this can be considered in some respect as a standard. It is not published by an international standardisation body but it acts somehow in the same way.
For me it is a bit a grey zone but if the document is intended to describe the structure in some formal way it may be considered as a valid entry.
@hoehrmann DCAT-AP has introduced foaf:page
for this:
https://semiceu.github.io/DCAT-AP/releases/3.0.0/#Distribution.documentation
I think https://www.w3.org/TR/2024/CR-vocab-dcat-3-20240118/ should be returned to the Working Group for further work.
While I can read in the document that I can use
dcterms:conformsTo
to reference the JSON Schema documentation for a JSON file, and I can usedcat:endpointDescription
to reference the documentation of an API, my purpose in reading through the document was to learn how to link a CSV file to a human-readable description of the structure of the file, or a JSON, XML, ... file in a custom format to a human-readable documentation of the structure, elements, attributes, properties, ... of the format. And I am coming up empty.If I ever had trouble finding the documentation for an API that I know exists, then
dcat:endpointDescription
might be useful to have. But I never really had such trouble. I do, however, very much have trouble working with proprietary CSV, XML, JSON, ... structures that are published without any documentation. And I feel very strongly that it must be obvious from this specification how to link data distributions to their documentation.