RE: Literals and URIs (2)

: Graham Klyne:
: Why should one not permit a URI with a conceptual mapping to a
: string of characters that can be used in different contexts to imply
: different actual meanings?
:
: I might define urn:mylexicon:Paris to refer to a lexicon entry containing
: all of the meanings you gave previously;  said URI might then reasonably be
: used as the object of a statement to reference one of those meanings.  I
: think this is an unarguably valid use of RDF (even though you may not like
: the style), which is very similar to the use of a literal "Paris" or a
: corresponding data URI discussed previously.   (I speculate that this could
: be similar to what happens with the DMOZ open directory description in RDF.)

This is perfectly resonable and seems the first step in a word sense
disambuigation process for RDF (the identificaton of possible
meanings/referents). Machine readable dictionaries with subject codes would be
one scalable approach to this. Something like DMOZ requires a bit more work from
the RDF application, although, DMOZ is available today.

It's interesting how directly one is moved into natural language processing when
considering literals.

Bill de hOra

Received on Thursday, 15 February 2001 20:00:00 UTC