RE: Re: Approximate matches in faceted schema implementations (PRODUCT FEATURES) WAS Introducing myself

>  So there you have it, metadata sandards (centrally controlled) are 
>  mandatory, as are controlled vocabularies.
>  
>  The question is - do faceted taxonomies help me achieve my aim more 
>  than topic maps, RDF, or other richer ontological markup?  Would the 
>  extensiblity of XFML create an uncontrolled chaotic environment in 
>  which many user taxonomies had to be mapped to many other user 
>  taxonomies (the n-squared problem)?  

XFML (the part that connects maps with <connect> and <psi>) assumes
that:
- Connections don't have to be perfect (better an imperfect system that
works) because:
- imposing metadata schemes centrally is often impossible

So the XFML approach of connecting metadata (distributed) may not be the
best way of handling it in an environment where you want central
control.

As for a comparison with RDF, XTM and such: XFML is ONLY a SIMPLE way of
exchanging a certain LIMITED type of faceted classification data. You
can happily use XTM (topicmaps) or RDF to store your faceted metadata
info. XFML is NOT the same thing as faceted classification, it is just a
simple format to EXCHANGE data. The question becomes: how complex a
datastructure do you want to have your metadata in, not which format do
you use to exchange it. XTM is optimized for one type of structure
(topicmaps). XFML is optimized for a very simple structure (hierarchical
faceted taxonomies). RDF is optimized for another structure (triples).
But what language you use is the wrong question. What structure you
want, that is the question. THEN decide on what technologies to use.
(Sorry for the capitals...)

>  Also, XFML promo blurb (and the XML.com article) suggests 
>  that two of 
>  the motivations behind XFML was that finding info was 
>  difficult as it 
>  was poorly structured and the search facilities were poor.  The 
>  implication being that because searching technology was so poor, it 
>  was necessary to define a new way of classification, not improve the 
>  way we search.  Is this understanding correct?  If so, why not 
>  address the searching problem first?

XFML can be used to improve search as well as browse. The whole
advantage of using taxonomies is that it structures your search (as in
the flamenco interface).
PeterV



----------
Thanks for playing. To unsubscribe from this group, send an email to:
[email protected]

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/