Broken URIs from http://registry.it.csiro.au/

w3c / dxwg

Data Catalog Vocabulary (DCAT)

https://w3c.github.io/dxwg/dcat/

Other

139 stars 55 forks source link

Broken URIs from http://registry.it.csiro.au/ #1517

Closed andrea-perego closed 1 year ago

andrea-perego commented 2 years ago

Following up from discussion in https://github.com/w3c/dxwg/pull/1510#issuecomment-1140330911

These URIs are used in examples of the DCAT3 specification.

To be decided if they need to be replaced.

riccardoAlbertoni commented 1 year ago

@dr-shorthair, @andrea-perego: Did we discovered any alternatives for the missing links pointed out in https://github.com/w3c/dxwg/pull/1510#issuecomment-1140415089?

dr-shorthair commented 1 year ago

Sorry - it was taken down a couple of years ago. No replacement :-( 'Linked data' eh?

rob-metalinkage commented 1 year ago

These are still broken - and as a reference to an ISO concept this represents a poor example. better to put an obvious stub to ISO - or even better an normative ISO endorsed URI even if its not operationally resolvable yet.

riccardoAlbertoni commented 1 year ago

@rob-metalinkage: I am afraid I am not correctly interpreting your suggestion.

Considering the codelist related to Role in EXAMPLE 53

ex:DS987
  a dcat:Dataset ;
  prov:qualifiedAttribution [
    a prov:Attribution ;
    prov:agent <https://www.ala.org.au/> ;
    dcat:hadRole <http://registry.it.csiro.au/def/isotc211/CI_RoleCode/distributor>
  ] ;
  prov:qualifiedAttribution [
    a prov:Attribution ;
    prov:agent <https://www.education.gov.au/> ;
    dcat:hadRole <http://registry.it.csiro.au/def/isotc211/CI_RoleCode/funder>
  ] ;
.`

what do you think could be a good replacement for http://registry.it.csiro.au/def/isotc211/CI_RoleCode/distributor?

the xml derived https://standards.iso.org/iso/19115/resources/Codelists/gml/CI_RoleCode.xml#distributor
the fictional http://iso.org/isotc211/CI_RoleCode/funder
other option I have missed

If you meant one of the first two options, I would consider to leave the URI as they are waiting for developments on the ISO side, or good proxies ;)

rob-metalinkage commented 1 year ago

I'd prefer a fictional but possibly future valid broken link than one guaranteed to be broken and confusing as a violation of principles of linked data. Either that or something like http://placeholderforiso.org

riccardoAlbertoni commented 1 year ago

Thanks @rob-metalinkage, I think I see your point and I appreciate your attempt to find a pragmatic solution. However, the more I think about it, the more I am not completely comfortable minting links in the ISO namespace. People might start using our fake ISO URI, and this might turn out to be more problematic than the issue we are trying to address.

On the second approach, I am slightly concerned about using a namespace that can be eventually hijacked by others, as in the case of http://placeholderforiso.org/

I suspect that doing nothing might be better than acting, and I propose moving this as a reminder for future addressing, hoping that in the future we will have time to find a more stable and definitive solution.

pchampin commented 1 year ago

Sorry for chiming in a little late, but I have another proposal: urn:example:isotc211/CI_RoleCode/distributor

this is a valid URN (see RFC6963 for the example namespace) and therefore a valid URI;
being a URN, it does not create wrong expectations about it being dereferenceable.

riccardoAlbertoni commented 1 year ago

Thanks, @pchampin I think this could be a compromise.
PR #1568 implements your proposal.