W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > September 2002

Re: Proposed N-Triples changes for datatypes & (untidy) literals

From: Patrick Stickler <patrick.stickler@nokia.com>
Date: Mon, 23 Sep 2002 15:10:56 +0300
Message-ID: <00a401c262fa$4532d320$d74416ac@NOE.Nokia.com>
To: "ext Jeremy Carroll" <jjc@hpl.hp.com>, "ext Jeremy Carroll" <jjc@hplb.hpl.hp.com>, "Dave Beckett" <dave.beckett@bristol.ac.uk>, "RDF Core" <w3c-rdfcore-wg@w3.org>


----- Original Message -----
From: "ext Jeremy Carroll" <jjc@hpl.hp.com>
To: "Patrick Stickler" <patrick.stickler@nokia.com>; "ext Jeremy Carroll" <jjc@hplb.hpl.hp.com>; "Dave Beckett"
<dave.beckett@bristol.ac.uk>; "RDF Core" <w3c-rdfcore-wg@w3.org>
Sent: 23 September, 2002 14:41
Subject: RE: Proposed N-Triples changes for datatypes & (untidy) literals


>
> Patrick:
> > I understood reification as requiring the specific untidy literal
> > node to be referencable. Is this not the case?
>
> Depends on your reading of reification, see:
>
> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Sep/0206.html
> Issue 4
>
>
> I worry that that is the thin end of a fat wedge, and so would rather not go
> there.
>
> Jeremy


Hmmmmm... are we then saying that reification captures the form
of expression rather than the meaning of expression? If so, then
rdf:subject, rdf:predicate, and rdf:object should take literals
rather than URIrefs as values. I.e.

   <rdf:subject>http://foo.com/blargh</rdf:subject>

and not

   <rdf:subject rdf:resource="http://foo.com/blargh"/>

But if the value of rdf:subject and rdf:predicate are going to be
the actual resources denoted in the original stating, then so
also should the value of rdf:object denote the actual resource
denoted in the original stating, and to do that, we either
have to add some extra machinery to how rdfs:range is interpreted
with regards to reifications, so that the same long-distance
datatyping applies to the reification as to the original triple,
e.g. add the following closure rule to the MT:

IF
   _:x rdf:type rdf:Statement .
   _:x rdf:predicate ppp .
   _:x rdf:object "LLL" .
   ppp rdfs:range ddd .
THEN
   _:x rdf:object ddd"LLL" .

[or something along those lines that results in the intepretation
of the object of rdf:object as I(ddd"LLL") etc.]

Or we have to be able to differentiate between and explicitly
reference the particular literal node of the reified triple.

I'd be happy with either. I think the latter would be easier,
and also offer the other legacy-related benefits as well.

Patrick

[Patrick Stickler, Nokia/Finland, (+358 50) 483 9453, patrick.stickler@nokia.com]
Received on Monday, 23 September 2002 08:15:18 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:51:02 EDT