Meeting minutes
RDF-star semantics: option 3 — https://github.com/w3c/rdf-star-wg/wiki/RDF%E2%80%90star-semantics:-option-3
enrico: proposal is to formalize full transparency
Souri: to be clear, we are not using the term "quoted triple" any longer?
enrico: I'm using random names for that concept
Option 3 does not allow triple terms in the subject position.
<AndyS> There are different ways to have that effect.
That is tantamount to literals in the subject position.
<AndyS> In semantics, already have literals as subject (rdfs3 rule in Semantics doc)
https://
An RDF triple consists of three components: the subject, which is an IRI or a blank node
That
That's like rdf:type "x" . Would be nice to prevent, of course.
<olaf> Unfortunately, I have to drop for a another meeting.
I *require* limiting triple terms to the object position, and *would like* to limit them to the "nameOf", but that has no precedent (cf. rdf:type "x"...).
Also +100 to Souri's point that triple terms adds complexity, let's add as little as is necessary.
The abstract syntax enforces things.
How is this different from generalized RDF?
https://
> Generalized RDF triples, graphs, and datasets differ from normative RDF triples, graphs, and datasets only by allowing IRIs, blank nodes and literals to appear in any position, i.e., as subject, predicate, object or graph names.
<doerthe> do we ignore the non wellformed parts of a graph or do we reject the whole graph?
Adding triple terms imply adding those to that mix. But it does not imply allowing triple terms as subjects.
Of course I see the philosophical "purity" of e.g. <<(<s> <p> <o>)>> a rdf:Triple . It's the same purity as "x" a rdf:Literal . I don't see the *practical value* of it; on the contrary.
Literals are practical things. Only blank nodes are "pure".
<enrico> rdf:type rdf:type rdf:type .
That is simply nominal nonsense.
<Souri> Simplify the proposed grammar?
<Souri> subject ::= iri | BlankNode | identifier object ::= iri | BlankNode | literal | identifier triple ::= identifier rdf:nameOf tripleTerm identifier ::= iri | BlankNode
I'll write that out line by line:
subject: := iri | BlankNode | identifier
object: := iri | BlankNode | literal | identifier
triple: := identifier rdf:nameOf tripleTerm
identifier: := iri | BlankNode
<tl> +1
that reduces to
subject: := identifier
object: := literal | identifier
triple: := identifier rdf:nameOf tripleTerm
identifier: := iri | BlankNode
<AndyS> Preference for "well-formedness" approach over abstract syntax with this situation coded into it.