– DRAFT –
RDF-star WG — Semantics TF

<william_vw> (unsure if that's needed here)

<tl> @william_vw we all do it

<enrico> we handle the queue here

discussion about creating instances of reificationProperty class

<niklasl> https://spinrdf.org

<AndyS> FWIW SPIN RDF is a dormant project and not active.

<doerthe> yes, right

<william_vw> niklasl yes indeed

<tl> :r rdfreifies <<( :s :p :o )>>. :r rdf:object [ .... ]

<niklasl> I think it could work it does imply that the reifier is a member of rdf:Statement, which we probably need some recommendation for.

<niklasl> Yes, as AndyS says, SHACL Rules would be valuable to look at.

<doerthe> was also my idea, but I put myself of the queue because I am reading up at the moment

<enrico> no, niklasl, reifiers are not statements!

<AndyS> OBJECT(triple term) = :o -- the function is on the tripleTerm not the reifier.

<doerthe> (after shacl inspection)

<niklasl> (A possible way using OWL infer that a specific reifier is also a classic reification: https://gist.githubusercontent.com/niklasl/69428b043be6f1d33fd45f89cbe52632/raw/13cd391a1eebad5f975e4aae93fd2fd581fb4683/statement-entailment.ttl - note: this is related to but *not* the same as "unstarring".)

<niklasl> (Something similar could be used for e.g. sh:TripleRule https://www.w3.org/TR/shacl-af/#rules .)

<tl> @niklas no

<william_vw> enrico thanks, that makes sense

<niklasl> Oh no, sorry, you're right tl :)

<Zakim> tl, you wanted to ask if the rdfs:domain of an rdf:ReificationProperty is ... always a reifier?

<niklasl> +1 (IIUC) that rdf:reifies is as strong as rdf:type (a core concept)

<niklasl> To note, I don't see much use for a reifier class...

<niklasl> Not even literals denote themselves (in 1.1)...

<enrico> fight FIGHT F I G H T !!!!!

<enrico> Example: https://agilemodeling.com/wp-content/uploads/2023/05/classDiagramAssociationClasses.gif

<enrico> Other example: https://images.slideplayer.com/24/6992546/slides/slide_3.jpg

<enrico> More example: https://vertabelo.com/blog/weak-entity-example/1.png

<enrico> UML/ORM/E-R

<niklasl> rdf:ReificationProperty rdfs:domain rdfs:Resource . # ?

<niklasl> I believe I agree that we need rdf:reifies, which is more detailed, requiring assertion and separate reification (in my words).

<william_vw> I'm also not convinced of this right now

<william_vw> (FWIW)

<doerthe> they are all grown-ups :D

<doerthe> to refine my position, I would like to have literals in subject position, but I understand why they could be dangerous. I still don't see the danger from triple terms

<niklasl> w3c/rdf-ucr#27

<gb> Issue 27 Integrating different ontology designs through entailment upon triple terms (by niklasl) [use case]

<niklasl> +1 for a less-generalized RDF (Sgt RDF)

<niklasl> R-Triples ?

<tl> G-Triples and LG-Triples

<niklasl> Input restriction sounds promising.

<niklasl> Postel's law recommends outputting input syntax (and accepting but not necessarily keeping other stuff).

<william_vw> I'm signing off (lunchtime)