Matthew Cockerill wrote:
> But from my perspective, the separation of concerns is a fantasic bonus, as it means that we can set
> up the identifiers in a certain way and stick to it. Meaning that we
can then accumulate lots of
> lovely semantic information over time (such an unambiguous metadata
telling us which book is which)
> and it remains valid.
Without weighing in on either side, I'd like to point out that one can
achieve the same thing using reference by description:
<rdf:Description>
<book:isbn>0393060349</book:isbn>
</rdf:Description>
Where the property book:isbn is defined to be an
Inverse-Functional-Property.
I can still accumulate metadata and there needn't be a single blessed
URI for either identification or location.
This approach has a lot going for it, especially where there can be
multiple URIs for identifying a resource. E.g. the info: URI that
ISBN and the URN.
Approaches such as RDF Forms [1] allow me to discover (by requesting
forms metadata from a service) useful things I can do with a book:isbn
predicate.
Cheers,
L.
[1].
http://www.markbaker.ca/2003/05/RDF-Forms/
Received on Tue Nov 15 2005 - 10:00:24 EST