[foaf-dev] [foaf-protocols] revisiting FOAF project goals

Olaf Hartig hartig at informatik.hu-berlin.de
Thu Jun 25 13:26:39 CEST 2009


On Thursday 25 June 2009 12:19:09 Matthew Rowe wrote:
> Hi Olaf
>
> Thanks for the comments.
>
> On 25 Jun 2009, at 06:14, Olaf Hartig wrote:
> > [...]
> > :Bob tv:trustworthiness [ rdf:type tv:TrustValue ;
> >                                       rdfs:value 0.9 ;
> >                                       tv:truster :Alice ]
>
> I like this form but I think there is an issue about how different
> people interpret trust. From a philosophical perspective when we say
> that we trust someone:
>
> Do we simply trust someone in general? i.e. In the broad sense of the
> term? Alice trusts Bob
> or
> Do we trust a person given a context? i.e. Alice trust Bob's opinion
> on the Semantic Web
>
> I am more inclined to say the latter.

Definitely agree. Trust is not only subjective but also contextual.

> Is there are any way this could be included in your vocabulary as something
> like: 
>
> Bob: tv:trustworthiness [rdf:type tv:TrustValue;
> 			rdfs:value 0.9 ;
> 			tv:truster :Alice;
> 			tv:context <http://dbpedia.org/resource/Semantic_Web> ]

My trust model for RDF statements does not consider context explicitly because 
I was assuming that the applications that provide implementations for the 
trust function know the context. Hence, the trust functions should already be 
contextual. For this reason the vocabulary -which was originally only meant 
for my trust model- does not contain something like a tv:context property. 
However, to be useful in the more general case the vocabulary should provide 
such a property. I will add it. Thanks for pointing this out.

Greetings,
Olaf


More information about the foaf-dev mailing list