W3C home > Mailing lists > Public > public-rdf-star@w3.org > August 2019

Re: RDF* semantics

From: Olaf Hartig <olaf.hartig@liu.se>
Date: Wed, 7 Aug 2019 09:13:28 +0000
To: "public-rdf-star@w3.org" <public-rdf-star@w3.org>
CC: Kingsley Idehen <kidehen@openlinksw.com>
Message-ID: <1565169206.8060.158.camel@lille71.ida.liu.se>
Hi Kingsley,

On Tue, 2019-08-06 at 12:16 -0400, Kingsley Idehen wrote:
> On 8/6/19 11:12 AM, Olaf Hartig wrote:
> > Hi Kingsley,
> >
> > On Mon, 2019-08-05 at 13:52 -0400, Kingsley Idehen wrote:
> >> [...]
> >> Under what circumstances in the real-world would the condition you model
> >> arise i.e., propositions that don't manifest as part of documentation?
> > For instance, we my want to capture that Alice told us that Bob's age is
> > 23, even if we don't have a document from Alice with this
> > statement/claim regarding Bob's age.
> 
> Through what medium did Alice make the aforementioned claim about Bob?

As I wrote, she *told* us (during a conversation in a cafe, if you
like).

> How is provenance articulated in said medium?

Similar to how you may have included provenance information in your
example document, Alice's spoken sentences may have included hints about
the provenance of her claim regarding Bob's age.

However, I don't see why this is relevant in our context in which we aim
to capture provenance of the claim regarding Bob's age in *another*
medium, where the provenance of the claim is that we have heard it from
Alice and the other medium is an RDF* graph (or a Turtle* serialization
thereof, if you like).

>  [...]
> >> The triple:
> >>
> >> @prefix : <#> .
> >>
> >> :bob foaf:age "23"^^xsd:integer . 
> >>
> >> Exists in a document identified by <> , so we a missing the following part
> >> of reality (as I understand it):
> >>
> >> @prefix : <#> . 
> >>
> >> <> a foaf:Document .
> >> <> :creator :i .
> >> <> :createdOn "2019-08-05"^^xsd:date . 
> >> <> foaf:primaryTopic :bob. 
> >>
> >> :bob foaf:age "23"^^xsd:integer . 
> >>
> >> Thus, it's the authoritative weight given to <>, by whoever, under whatever
> >> situation and circumstance, that leads to acceptance or rejection of the
> >> claims outlined e.g., those about the foaf:age of the entity identified
> >> by :bob .
> > I am not questioning that. Note, however, that your document is not
> > making a statement/claim about the claim regarding Bob's age (other than
> > asserting it, of course).
> 
> The document is the medium through which those claims are projected. Put
> differently, it provides the surface for the claim represented as
> sentences/statements.

I have no problem with that. But again, while your example document
contains the statement about Bob's age and some statement about itself
(including provenance-related statements), it does not contain any
statement about the statement about Bob's age.

> > The purpose of RDF*, and of RDF reification,
> > is to allow you to do so (within your document or within another of your
> > documents).
> 
> I understand the RDF reification part since it allows one speak directly
> about units of communication or "parts of speach" (subject, predicate,
> and object components of RDF sentences/statements) using documentation,
> just like anything else.
> 
> At this juncture, I see RDF* as being about an algorithm (delivered as
> syntax sugar) for attempting to address the verbosity of  existing RDF
> reification syntax rather than dealing with new RDF semantics.

Fine with me. Under the perspective that considers RDF* purely as
syntactic sugar, we have the existing RDF semantics and don't need
anything new (except for a mapping that defines how the syntactic
extension maps to standard RDF, but that we have).

However, as an alternative to the syntactic-sugar perspective, another
perspective is to consider RDF* to be based on its own abstract data
model (which is an extension of the RDF data model). Then, as has been
done for the RDF data model, we may define a semantics related to the
RDF* data model. Pat's initial point was that such a semantics will help
to provide some clarity (under this alternative perspective, I assume).

Olaf

> Anyway, that's why "RDF* semantics" is confusing to me with regards to
> its fundamental objective at this point in time.
> 
> Kingsley
> 
> >
> > Best,
> > Olaf
> >
> >> If we get the world we a modeling for clear, other bits of the problem
> >> will become clearer.
> >>
> >> Conclusion:
> >>
> >> In my opinion, we need a complete example that's relatable to the world
> >> that we are describing using RDF sentences :)
> >>
> >
> 


Received on Wednesday, 7 August 2019 09:13:57 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:02:56 UTC