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

Olaf Hartig hartig at informatik.hu-berlin.de
Fri Jun 26 07:19:04 CEST 2009


Hey Peter,

On Thursday 25 June 2009 23:51:32 Peter Ansell wrote:
> 2009/6/26 Dave Brondsema <dave at brondsema.net>:
> > Matthew Rowe wrote:
> >> [...]
> >> Bob: tv:trustworthiness [rdf:type tv:TrustValue;
> >>                                       rdfs:value 0.9 ;
> >>                                       tv:truster :Alice;
> >>                                       tv:context
> >> <http://dbpedia.org/resource/Semantic_Web> ]
> >
> > With Konfidi's vocabulary http://www.konfidi.org/ns/trust/1.2# I do:
> >
> > [ a k:Relationship ;
> >    k:truster <http://brondsema.net/dave#> ;
> >    k:trusted _:andy ;
> >    k:about [ a k:Item ;
> >        k:topic ktopic:internet-communication ;
> >        k:rating 0.95
> >    ]
> > ].
>
> If you assume that people will create a single relationship for each
> trusted topic it could be much simpler than that.
>
> :trustAndyInternet a k:TrustRelationship ;
>
>      k:truster <http://brondsema.net/dave#> ;
>      k:trusted _:andy ;
>      k:topic ktopic:internet-communication ;
>      k:rating 0.95 .
>
> Relying on people understanding the completely abstract k:about and
> k:Item isn't useful if they can just as easily create the relationship
> with a single level of structure and have the same functionality in
> the end.

Agree.

However, do you assume people will publicly publish their trust statements at 
all? I doubt that, mainly for privacy reasons. Maybe once people can apply a 
proper authorization mechanism for their private data we could expect trust 
statements. Until then the trust statements are inside the applications. I 
don't see no reason to materialize these statements as RDF data unless the 
application stores all its data as RDF anyways. Dave, what is the reason to 
represent the relationships in Konfidi with RDF?

Greetings,
Olaf


More information about the foaf-dev mailing list