Re: [Patterns] Materialize Inferences (was Re: Triple materialization at publisher level)

Giovanni Tummarello wrote:
> The problem is not the standard but its the process.
> 
> a webmaster should dump the data in a owl or so reasoner (or rdfs
> whatever) , then pick the outcoming triples and basically hand write
> those extra classes in the nice HTML template.. mmm :-)

I don't think full inference is necessary or feasible. Each publisher
should consider which triples are likely to be useful for their clients,
and materialize those. Obviously there's no use writing:

	X rdf:type rdfs:Resource ; owl:sameAs X .

But it makes a lot of sense to write:

	X rdf:type sioct:BlogPost , sioc:Post , sioc:Item .

And it shouldn't be difficult for webmasters who understand the basics
of RDF and the semantics of the vocabularies involved. Which they
should :-)

By the way, it would be nice if the description of this design pattern
included more guidelines/examples of which things are generally useful
to materialize and which are not.

-- 
Vasiliy Faronov

Received on Wednesday, 7 April 2010 11:22:44 UTC