Closed bertvannuffelen closed 9 months ago
As I understand Bulk download requirement mostly applies to Data Services, because Data Services might provide row by row access to data, but not Bulk download.
If we have:
ex:dataset1 a dcat:Dataset ;
dcat:distribution ex:dist1 .
ex:service1 a dcat:DataService ;
dcat:servesDataset ex:dataset1 .
ex:dist1 a dcat:Distribution ;
dcat:accessService ex:service1 .
and ex:service1
does not support Bluk download. Then how we can tell, that ex:dist1
has a single downloadable file?
@sirex I view the requirement basically as having both a dataservice and a downloadable file as distributions of an HVD dataset:
ex:dataset1 a dcat:Dataset ;
dcat:distribution ex:dist1 , ex:dist2 .
ex:dist1 a dcat:Distribution ;
dcat:accessService ex:service1 .
ex:service1 a dcat:DataService ;
dcat:servesDataset ex:dataset1 .
ex:dist2 a dcat:Distribution ;
dcat:accessURL <file> ;
dcat:downloadURL <file> .
So basically for HVD dcat:downloadURL
becomes «mandatory»? Now it is «optional».
Well, it is optional to support distributions which are not directly downloadable, e.g., according to DGA. But for directly downloadable files, it should be used. And therefore, for HVD dataset's distributions in the form of bulk download, this should be mandatory.
The feedback from @jakubklimek is correct.
In case the dataset is subject to the HVD IR, and the HVD IR mandates the presence of a bulk download (almost always the case), then the downloadURL is mandatory.
The HVD implementing regulation requires to provide bulk downloads for datasets. Bulk download is a functionality that allows to get the whole dataset as a copy on your local computer.
proposal
As in DCAT-AP guidelines is specified, a downloadable file is to be represented with a Distribution.