Re: Event Updated: RDF-star WG biweekly meeting

Am 29.08.2024 um 15:27 schrieb Franconi Enrico <franconi@inf.unibz.it>:

On 29 Aug 2024, at 13:13, Doerthe Arndt <doerthe.arndt@tu-dresden.de> wrote:

Also, in this way Pat & al. (possibly unwillingly) made a perfect correspondence with F-Logic by Michael Kifer. This also allows for the full equivalence with a first order approach where triples have the semantics of a single ternary T/3 predicate.

These two things I knew, but in my opinion the double mapping, i.e. applying IEXT after IS is enough to get there.

I guess that you are right, as soon as the double mapping “passes through" resources.

This by the way even allows quantification over the predicate since we quantify over  IR and not over 2^(IPxIP) which would lead to higher order.

Note that in RDF semantics any property is also a resource IP ⊆ IR, due to the meta modelling rule:

               u a y
---------------------
a rdf:type rdf:Property

So, it seems that IP is totally irrelevant.


Not totally irrelevant, it makes the semantic condition for RDF semantics easier to state (source: https://www.w3.org/TR/rdf11-mt/#rdf-interpretations)

x is in IP if and only if <x, I(rdf:Property)> is in IEXT(I(rdf:type))

So, RDF semantics does not rely on a construct as you needed it for graph terms which would be


[I+A](t) = TRUE   implies
          <[I+A](t.p), [I+A](rdf:Property)> ∈ IEXT([I+A](rdf:type))


this is basically your entailment rule from above but on semantic level. So, I finally see that it is nice that the class of properties in RDF corresponds to a concrete set which even has a name. However, as the above definition makes that IP=ICEXT(I(rdf:Property)) if we are in RDFS notation, you could argue that the advantage is rather a matter of taste.


Kind regards,

Dörthe



I see how we both like these kinds of conversations ;)

Oh yeah!

—e.