Re: Reviving Semantic Relationships in Beagle



Hi David, hi Kevin,

David Peterson wrote:
Are you aware of the Aperture project [1]? They have developed a crawler and
RDF extractor that crawls the filesystem as well as the web. Very
comprehensive system. I believe this project could utilise aspects of the
Aperture system. Beagle already does the crawling so you could use the
getRDFContainer method (which just takes a URI) and spits out an RDF model.

You can then use this RDF model in whatever way you would want inside of
Beagle.
I think the current step of semantifying Begale (adding RDF capabilities) is yet at a simpler level. The idea is first to only provide a basic RDF query interface on top of the existing infrastructure and Lucene indices. Just to open the stored metadata to RDF clients, just to see how this performs and what it enables. As far as I can see, this issue is feasable.

Kevin Kubasik wrote:
But, in my opinion, the best improvement
would be to arrange beagle metadata in a OWL ontology.
I'm sorry, most of us are pretty new to the whole RDF terminology
thing, I've heard of ontology, but OWL?
Well, in simple words, on ontology can be defined using RDFS (RDF Scheme) or OWL. OWL is more powerful, but RDFS is sufficient for current RDF activities regarding Beagle. The Beagle ontology basically should give URIs to all kinds of metadata it knows about (this are the field names of the Lucene Documents).

In this way, metadata could populate a Knowledge Base of system.
Clearly, this involve a preexisting OWL rules.
The challenge, for me, is to build a GUI that easily allow Description
Logic search.
Building on before, I'm a little lost by what your getting at here,
could you explain in a little more detail, or give an example? Thanks!
I think reasoning and DL is two or three levels above the current first step.

Cheers,
Enrico

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]