W3C home > Mailing lists > Public > semantic-web@w3.org > March 2008

Re: meta-information about assertions

From: Phil Archer <parcher@icra.org>
Date: Mon, 03 Mar 2008 14:24:54 +0000
Message-ID: <47CC0A36.5050408@icra.org>
To: semantic-web@w3.org

This is a question that has been around for more than a little while. 
Reading the specs, you'd say that reification was the answer but this is 
often regarded as "a dangerous thing" and the advice seems to be to 
steer clear of it if possible. Named Graphs have been proposed (and well 
defined) but have not been incorporated into the core specs (although 
SPARQL offers some support).

So the safe answer for now is to include some triples that have the RDF 
instance itself as the subject and only merge the triples in the 
instance as a whole if the self-description leads you to trust it.

So you might see

<> foaf:maker <http://www.fosi.org/people/philarcher/foaf.rdf#me>
s1 p1 01
s2 p2 o2

and decide to trust it but then see a different instance with

<> foaf:maker <http://bad.example.org/foaf.rdf#me>
s1 p1 01
s2 p2 o2

and decide not to trust that data and ignore it.

For now, as I understand it, that's it. If you want more, you'll need to 
wait for the RIF - and even then you can't really write a rule that says 
"I met this chap in the pub and I like him so I'll trust what he says" - 
that's a matter of human judgement.

This post in the context of POWDER which has been wrestling with this 
for a good while - new docs coming out v. soon with our solutions based 
on this.

Phil.


Cristiano Longo wrote:
> Of course. I seen also Quadruples.
> 
> --- Frank Manola <fmanola@acm.org> ha scritto:
> 
>> Have you looked at
>> http://www.w3.org/TR/swbp-n-aryRelations/ and the  
>> work on named graphs (and provenance in general)?
>>
>>
>> On Mar 3, 2008, at 4:45 AM, Cristiano Longo wrote:
>>
>>> Hi all,
>>> i'm trying to merge rdf(more specifically OWL)
>> graphs
>>> from different sources using collaborative
>> filtering
>>> and trust related technologies. But my question
>> is:
>>> what is the proper way to encode a "meta
>> assertion"
>>> like "A says X about B", in order to deal with
>>> contraddictory assertions?
>>>
>>> Reification? Using SKOS? Something else?
>>>
>>> Thank you in advance.
>>>
> 
> 
> 
>       ___________________________________ 
> L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail: http://it.docs.yahoo.com/nowyoucan.html
> 
> 
> 

-- 
Phil Archer
Chief Technical Officer,
Family Online Safety Institute
w. http://www.fosi.org/people/philarcher/
Received on Monday, 3 March 2008 14:25:10 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:41:09 UTC