RE: Faceted Search Interface

Just to follow up on Danny's comment: our approach is metadata-based by
default, but can be used as Danny states below to implement faceted search.
The demo Mark links to is not very faceted in the traditional sense that I
believe John is speaking of. We're finding that in practice, there is a
continuum from simple Dublin Core type metadata search to truly faceted
applications, and that, as Danny says, representing range distinctions in
the facets helps to bring the faceted aspects of a given application to the
fore.

- regards, BPA

Bradley P. Allen
Founder and CEO
Siderean Software LLC
5155 West Rosecrans Avenue, Suite 1078
Los Angeles, CA 90250
Phone +1 310 491-3424
Fax +1 310 379-0231
Web www.siderean.com


> -----Original Message-----
> From: Danny Ayers [mailto:danny666@virgilio.it]
> Sent: Wednesday, June 18, 2003 6:24 AM
> To: John S. Erickson; Butler, Mark; www-rdf-dspace@w3.org
> Cc: bpa33@bpallen.com
> Subject: RE: Faceted Search Interface
>
>
>
> > The associated white paper is useful:
> > http://www.siderean.com/TechnologyWhitePaper.pdf
> >
> > ...although one realizes that this "faceted search" is simply
> > "metadata-based
> > search..."
>
> I raised this point on the xfml list very recently, and got a very good
> response from bpallen at Siderean. They are representing facets by
> specifying the range of the properties (with RDFS). This, he
> agreed, didn't
> formally incorporate the disjointedness characteristic of faceted metadata
> (which would need owl:disjointWith or equivalent) but that in practice
> lexical distinctions in the URIs enforced the disjointness. So I think to
> all practical intents and purposes, it is fair to talk of this as being
> faceted search.
>
> Cheers,
> Danny.
>
> ----
>
> http://dannyayers.com
>

Received on Wednesday, 18 June 2003 10:20:12 UTC