Meeting minutes
Discussion on PR #91 : A formal background to unify triples and triple terms
Discussion about the PR, whether RE could be the identity or not.
Discussion about how model theory makes RDF scary for many people. The RDF Semantics should aim to be less scary.
Discussion about the term "fact" , and the importance to frame it in the context of an interpretation.
No objection among the native participants to merge the PR with the suggested changes.
Discussion on issue #49 : Define an interpretation of Triple Terms
niklasl: this issue led to discuss the relevance of the sections about reification, containers... in the semantics document
… Enrico was worried about my proposal to "expand" the structure of triple terms.
… Triple terms could "pop out" of the spurious use of the corresponding properties.
ack
pchampin: I think the sections in RDF-SEMANTICS about reification, containers... are comments on those terms from the perspective of semantics.
… They are not defining those terms. RDF-SCHEMA is. This could perhaps be made more clear.
… For an entailment regime that "expands" triple terms, this should be at most a note.
niklasl: I generally agree. Note that this entailment regime would make unstar moot.
… I'll work on making my case.
pchampin: the 'unstar' mapping manages to roundtrip by "reserving" the vocabulary used to expand.
… If we make them part of entailment, we can not forbid their use.
niklasl: I sympathize with that. But this "reserved" vocabulary could maybe be less "private".
… Compare with the rdf:langString datatype, which you are not supposed to use directly.
Discussion on issue #61 : Explain how classic RDF reification relates to triple terms and rdf:reifies
niklasl: linked to the previous issue.
… In a previous conversation, people suggested that old-style reification nodes were a kind of reifier.
AndyS: should we say *in what way* they are a kind of reifier?
… Is the new thing is a superset of the old thing? A replacement for it?
… As soon as you say there is a relationship, you imply that the old stuff does not need to be replaced.
niklasl: yes, these are important points.
… The class rdf:Statement is a set, conceptually a subset of reifiers.
… But you can't infer that there is a reified triple term from an instance of rdf:Statement (its properties can be broken, incomplete...).
… You could infer the other way around.
AndyS: we need this in a document to discuss it further.
niklasl: my hope is that this could be included in the note.
pchampin: +1 to have this explainations, but in rdf-primer, not rdf-semantics
… the new style reification has two advantages compared to old style: more expressive (marriages, etc.), and atomic
tl: I don't think that the primer should mention old-style reification
… OTOH, people for whom old style reification does not the job should not need to change.
… For me, old style reification is exactly the same thing as rdf:reifies with exactly one triple term.
niklasl: I agree with tl that we should not make any of our documents too much about old-style reification.
pchampin: I disagree with tl about the primer not mentioning old-style reification.
… The primer is also a place where RDF 1.1 user will look for info about what's new.