In W3C DCAT data files are specified in the dataset metadata distribution list:
You can have extra metadata (extendible beyond DCAT if required) on a per file basis.
All id (@id) should be permanent identifiers.
DATA ACCESS
distribution id (@id) should return the distribution metadata. (convenience)
API access through "accessURL".
Web clients (e.g. HTTP GET) access through "downloadURL".
Note: The IRIs could redirect to the resource, wherever it is stored (in case of data files probably an Object Store).
Related to FAIR data. (We are already done it!) Datasets point to data files in their distribution list.
Good rationale here
In W3C DCAT data files are specified in the dataset metadata distribution list: You can have extra metadata (extendible beyond DCAT if required) on a per file basis. All id (@id) should be permanent identifiers.
DATA ACCESS
distribution id (@id) should return the distribution metadata. (convenience)
API access through "accessURL".
Web clients (e.g. HTTP GET) access through "downloadURL".
Note: The IRIs could redirect to the resource, wherever it is stored (in case of data files probably an Object Store).