Re: [dxwg] Is one of dcat:accessURL, dcat:accessService and dcat:downloadURL required in a dcat:Distribution? (#1197)

Thanks for your explanation and contribution, @nfreire . 

I would say that your option 2 (i.e., the one using `dcat:accessService`) is more in line and conformant with DCAT 2. 

About the pattern `prov:wasGeneratedBy` / `schema:SearchAction` / `prov:used` (which is not in DCAT 2):
- IMO, this is an interesting approach that can be potentially applied also to other service types / protocols. Actually, you are outlining a solution for an issue that was raised in the collected use cases (see, e.g., https://www.w3.org/TR/dcat-ucr/#ID18), but that it was not addressed in DCAT 2.
- I wonder whether you considered using instead `prov:wasDerivedFrom`, possibly in its qualified form (so to link together the distribution, the data service and the search action with a ternary relationship). 

An alternative option would of course be to instantiate the SPARQL endpoint URL with the query itself, and use it as the `dcat:downloadURL` of the distribution. 

-- 
GitHub Notification of comment by andrea-perego
Please view or discuss this issue at https://github.com/w3c/dxwg/issues/1197#issuecomment-571827393 using your GitHub account

Received on Tuesday, 7 January 2020 23:51:49 UTC