Re: My opinion of today on the profiles proposal

Hi Olaf,

On Fri, Apr 26, 2024 at 10:42 AM Olaf Hartig <olaf.hartig@liu.se> wrote:
>
> Thanks for the confirmation Enrico!
>
> I am a bit surprised by this observation. I was always assuming that
> the definition of the semantics would include defining what the
> property denoted by the IRI rdf:reifies means. Seeing now that this is
> not the case, I wonder: do we have any definition of what this property
> means?

Some definitions or explanations floating around in emails, but
nothing definitive in a document yet.

> The reason why I ask is that I am still trying to get my head around:
>
> i) what exactly the kind of thing denoted by IRI :r in a triple such as
> the following is:
>
>  ( :r , rdf:reifies , ( :s, :p, :o ) )

It is anything which reifies--as in concretizes--an abstract
relationship; that relation being a simple logical expression as
encoded by a triple.

We're calling these things "reifiers" (thus labelling their *role* as
subjects of an rdf:reifes triple, not their nature). For examples, see
e.g. [1] and [2].

> ... and ii) what exactly the relationship between the thing denoted by
> IRI :r and the triple term in that triple is, and also

Suggested definition:

    rdf:reifies a rdf:Property ;
        rdfs:comment "A property relating anything more concrete or
specific to one or more abstract relationships."@en ;
        rdfs:domain rdfs:Resource ;
        rdfs:range rdf:Triple .

> iii) what the relationship between the statement captured by the
> following triple and the triple term in the previous triple is (if
> any)?
>
>  ( :r , :p2 , :o2 )

Formally, the same kind of relationship as between any objects that
are related to by the same subject. I.e. it depends on the nature of
these relationships when interpreted together.

If the reifer :r is an :Assertion :madeBy and agent :atDate sometime,
then it's rdf:reifies relation is likely defined, for the :Assertion
class, to be restricted with cardinality 1, and it represents an
assertion of that abstract relationship denoted by <<(:s :p :o )>> at
sometime by said agent. If it is a :Marriage as a kind of situation,
it is the more concrete circumstance that is, expressed using
rdf:reifies (the relationship, not the PName), related to likely many
abstract relationships, that can be seen as "originating from" or
having been "formed by" that concrete thing. (Concrete here meaning
the marriage when conceptualized as a concrete, endurant situation,
with an originating date and location (or related to the ceremony
reifying those abstract properties), and causing people to become
spouses, being related to each other, vows to be made, a deity to
forge (ostensibly lifetime-delimited) bonds of souls, etc.)

Best regards,
Niklas

[1]: <https://lists.w3.org/Archives/Public/public-rdf-star-wg/2024Apr/0158.html>
[2]: <https://lists.w3.org/Archives/Public/public-rdf-star-wg/2024Apr/0167.html>


> Olaf
>
>
> On Fri, 2024-04-26 at 08:05 +0000, Franconi Enrico wrote:
> > On 26 Apr 2024, at 09:14, Olaf Hartig <olaf.hartig@liu.se> wrote:
> > > Looking through them now, and also looking again at the definition
> > > of
> > > the semantics of Profile 1, as given in
> > >
> > > https://github.com/w3c/rdf-star-wg/wiki/RDF%E2%80%90star-semantics:-option-3#semantics
> > >
> > > it seems to me that the IRI rdf:reifies is in no way different from
> > > any
> > > other IRI (from the semantics perspective). In other words, the
> > > meaning
> > > of the property denoted by this IRI does not seem to be defined as
> > > part
> > > of this definition. Is this observation correct?
> >
> > Yes. If you see at the syntax of the well formed fragment
> >
> > graph      ::= ( triple | reifier rdf:reifies tripleTerm )*
> > triple     ::= subject predicate object
> > subject    ::= iri | BlankNode
> > predicate  ::= iri_but_rdf:reifies
> > object     ::= iri | BlankNode | literal
> > tripleTerm ::= triple
> > reifier    ::= iri | BlankNode
> >
> > the use of rdf:reifies property is severely restricted in the syntax,
> > but its denotation is unrestricted as a property, namely it behaves
> > like any other property, e.g., it is many-to-many.
> > —e.
> >

Received on Saturday, 27 April 2024 08:34:11 UTC