search

ICA-EGAD / RiC-O

ICA Records in Contexts-Ontology (ICA RiC-O) GitHub repository web pages
https://ica-egad.github.io/RiC-O/
50 stars 17 forks source link

End ontology IRI with a hash (#) #36

Closed ivozandhuis closed 1 year ago

ivozandhuis commented 2 years ago

I think the IRI of rico should end with a hash ('#"). So https://www.ica.org/standards/RiC/ontology# instead of https://www.ica.org/standards/RiC/ontology.

andreasnef commented 1 year ago

Hi @ivozandhuis

I agree, that's why I created a PR quite a while ago already: #31

ivozandhuis commented 1 year ago

Background of this issue is that ignorant users could cut-and-paste the name of the RiC-ontology (https://www.ica.org/standards/RiC/ontology) and use this as a prefix (must be: https://www.ica.org/standards/RiC/ontology#). If we make the name of the ontology the same as the prefix, this error is made less often.

I've studied some examples of names of ontologies and whether the ending character is part of the name or not. There seems to be no real best practice here:

  1. http://www.w3.org/1999/02/22-rdf-syntax-ns#
  2. http://www.w3.org/2000/01/rdf-schema#
  3. http://www.w3.org/2002/07/owl
  4. http://www.w3.org/2004/02/skos/core
  5. http://www.w3.org/ns/prov#
  6. http://www.loc.gov/premis/rdf/v3/
  7. http://id.loc.gov/ontologies/bibframe/
  8. https://schema.org/

In cases 1, 2, 5, 6, 7, and 8 the name and the prefix are the same. In cases 3 and 4 (and those are not the least) not. Still, based on this sample I would suggest to add the '#' sign to the name of the ontology.

florenceclavaud commented 1 year ago

Hi @ivozandhuis, thanks a lot for this review.

In the list you have made, there are vocabularies that use "hash namespace" and others that use "slash namespace". RiC-O has used "hash namespace" from the beginning - I chose this because I wanted it to remain quite simple to handle.

Just to explain why the IRI of the ontology has no 'hash' (https://www.ica.org/standards/RiC/ontology), unlike the namespace: I have followed recipe 3 as defined in the W3C Best Practice Recipes for Publishing RDF Vocabularies. See also, about all this, the Appendix B: URI namespaces of the same document. Then I asked the company that hosts the ICA website to configure the server used, so that the RiC-O URI can be dereferenced just as this document recommends it should be in such a situation. It took time, many emails and iterative tests, to get the result expected (and I will redo the tests soon, since more than two years have passed and things may not work so well any more now.) Also, like any RDF resource, an ontology should not change IRI - at least if there is a change, it should be done for a very strong reason and be documented. Of course we still have a draft only now, and we could consider it is still time to change it. Besides, it would not affect the IRIs of the components of the ontology. But we also have to consider that, if we change the IRI, I will have to ask the hosting company that they change the configuration, then to test everything, and also to make sure that 'https://www.ica.org/standards/RiC/ontology' is redirected to the new IRI. This said, I do not mean at all that we should close this issue just now and that the PR @andreasnef created should be rejected based on these facts. On the contrary I think we have to investigate and think a bit more on this before making a decision. I have checked other resources and want to go a bit further.

Any other thoughts on this everybody?

jbkrause commented 1 year ago

Summary:

Pragmatic suggestion:

As at least one W3C recommendation does not include the hash and this is applied in two major ontologies. There is no imperial reason to append an # to the URI. Therefore, I suggest to go for the simplest and least impacting solution, i.e. to keep the ontology URI as it is in future RiC-O versions.

florenceclavaud commented 1 year ago

Thanks @jbkrause ! I agree. And so, as you, Ivo and I have reached a consensus about this, I am closing this issue and will also close the pull request without merging.

dpitti commented 1 year ago

Not that you need me, but I agree. And good sleuthing!

Daniel

From: Florence Clavaud @.> Sent: Wednesday, August 23, 2023 11:01:36 AM To: ICA-EGAD/RiC-O @.> Cc: Subscribed @.***> Subject: Re: [ICA-EGAD/RiC-O] End ontology IRI with a hash (#) (Issue #36)

Thanks @jbkrausehttps://github.com/jbkrause ! I agree. And so, as you, Ivo and I have reached a consensus about this, I am closing this issue and will also close the pull request without merging.

— Reply to this email directly, view it on GitHubhttps://github.com/ICA-EGAD/RiC-O/issues/36#issuecomment-1690127285, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAMUL7ZLNCIPW2HHBAZJFLDXWYLNBANCNFSM5UK62OVA. You are receiving this because you are subscribed to this thread.Message ID: @.***>