Re: SPARQL-WG comment on rdf:text

> Given that the relevant values already exist as interpretations of
> plain literals in any flavor of RDF interpretation, and could easily
> be made to live in the value space of rdf:text, is there any
> reason to ever use a literal "fou@fr"^^rdf:text to denote such a value?

As I understand it, RIF needs it.  The RIF language specs are written to
be independant of RDF.  The syntax and semantics don't know anything
about RDF Plain Literals or about rdf:text, except to say that support
for rdf:text is a required datatype.  Then, in "RIF RDF and OWL
Compatibility" [1], we give the mapping between RDF plain literals
(which the rest of RIF doesn't know about) and rdf:text literals (which
is just another datatype).

This is where rdf:text came from, then called rif:text.  But it turned
out OWL 2 wanted to do the same thing, so we factored it out.  (Let's
not get into second-guessing, at this point, whether that was a good
idea.)

I don't doubt that RIF could be rewritten to use RDF plain literals
instead of rdf:text, but that certainly doesn't seem like the easiest
path forward at this point.

In my mind, the only difficult question is whether some people *want* to
use rdf:text with SPARQL.  I suggest we just say no, for now, by saying
that rdf:text MUST NOT be used in any language, format, protocol, or API
that does not explicitely permit rdf:text to be used.  Then, some day,
in some version of SPARQL, they can add a flag or trigger to allow it,
or something, if they want.

      -- Sandro

[1] http://www.w3.org/2005/rules/wiki/SWC

Received on Saturday, 16 May 2009 00:19:23 UTC