Define a URI for the class of embedded triples for use with rdfs:range and rdfs:domain.

w3c / rdf-star

RDF-star specification

https://w3c.github.io/rdf-star/

Other

119 stars 23 forks source link

Define a URI for the class of embedded triples for use with rdfs:range and rdfs:domain. #102

Closed afs closed 2 years ago

afs commented 3 years ago

In order to support the use of rdfs:range and rdfs:domain, the spec should have vocabulary defined for the class of all embedded triples. Specifically, a standard URI for the class.

hartig commented 3 years ago

Is the idea to simply introduce such a URI without explicitly defining any specific semantics for it, or should it also have a specific semantics? For instance, assuming rdfstar:EmbTriple would be this URI, should the following assertion lead to any special inference? If yes, what inference?

:arbitraryURI rdf:type rdfstar:EmbTriple .

I mean, the URI :arbitraryURI clearly is not an embedded triple.

afs commented 3 years ago

This is the same situation as rdfs:Literal.

c.f. :arbitraryURI rdf:type rdfs:Literal.

No entailments.

It should go into rdf: - other work has put URis into that namespace (e.g. rdf:PlainLiteral, rdf:JSON).

Later migration of URIs is extremely problematic on the web. c.f. We are stuck with application/x-www-form-urlencoded. Nowadays good practice is to use the unregistered but intended final name for a MIME type.

hartig commented 3 years ago

Makes sense. Thanks for the clarification.

gatemezing commented 3 years ago

Sorry @afs but I don't get it. Do you mean to update the rdf vocab at https://www.w3.org/1999/02/22-rdf-syntax-ns#, and add sth likerdf:embTriple? Wouldn't be better to have a separate namespace of a vocab with specific axioms ?

afs commented 3 years ago

(As "Literal" is in RDFS, this should be as well, Never can remember the split ...)

Just write in the RDF-star report "we define rdfs:TripleTerm as the class of all RDF-star embedded triples" (or whatever name) and include like:

rdfs:Literal a rdfs:Class ;
    rdfs:isDefinedBy <http://www.w3.org/2000/01/rdf-schema#> ;
    rdfs:label "Literal" ;
    rdfs:comment "The class of literal values, eg. textual strings and integers." ;
    rdfs:subClassOf rdfs:Resource .

RDF-star is a proposal for a new feature to RDF and any future WG would incorporate it.

A separate namespace is effectively permanent. Like the RDF/RDFS split is (hindsight) unnecessary, a third namespace is confusing.

gatemezing commented 3 years ago

OK. That's more like a suggestion to add in the draft. In that case, if we have

rdfs:TripleTerm a rdfs:Class ;
    rdfs:isDefinedBy <http://www.w3.org/2000/01/rdf-schema#> ;
    rdfs:label "Triple Term"@en ;
    rdfs:comment "The class of triple terms for rdf-star."@en ;
    rdfs:subClassOf rdfs:Resource .

Aren't we forcing to entail that the triple term is a rdfs:Resource? Wouldn't this have other consequences per your comment about no entailments? Oh wait, Maybe you meant this definition below?

rdf:TripleTerm a rdfs:Datatype ;
    rdfs:label "Triple tem"@en ;
    rdfs:comment "The class of all RDF-star embedded triples."@en ;
    rdfs:subClassOf rdfs:Literal ;
    rdfs:isDefinedBy <http://www.w3.org/1999/02/22-rdf-syntax-ns#> ;
    rdfs:seeAlso <https://w3c.github.io/rdf-star/cg-spec/editors_draft.html#dfn-triple> .

BTW, who are the editors of the rdfs vocab? They updated the vocab in 2019 without proper mention of metadata? Where are the best practices in the W3C namespaces? :)

afs commented 3 years ago

While it behaves like a literal, (not deferenced) it isn't a subclass. It is a new RDF term along side the others. Literals can't go in the subject position.