Re: PROPOSAL to close ISSUE-39: RDFa term mapping triples

On Tue, 5 Oct 2010 15:46:57 +0200
Ivan Herman <> wrote:

> I did (and still do to a certain extend) have my issues, yes, so
> indeed it is unfair to put the weight on solely Mark's (admittingly
> wide) shoulders:-). I think Toby had similar issues, too.

Indeed - I think Ivan's original way of modelling terms and prefixes is
the way to go. Richards proposal is simpler, sure, but profiles are a
WORM format (write once, read many -- in this case, use many).

I was prepared to go with Richard's proposal in the interests of
consensus when it looked like consensus was heading in that direction,
but I do feel that Ivan's existing vocab is the way to go.

In fact, I'd like to add a couple of extra terms to it:
rdfa:ProfileDocument (a subclass of foaf:Document) and rdfa:defines. So
that the profile graph looks like:

	  a rdfa:ProfileDocument ;
	  rdfa:defines [
	    a rdfa:TermMapping ;
	    rdfa:term "title"^^xsd:string ;
	    rdfa:uri ""^^xsd:anyURI
	  ] ;
	  rdfa:defines [
	    a rdfa:PrefixMapping ;
	    rdfa:prefix "foaf"^^xsd:string ;
	    rdfa:uri ""^^xsd:anyURI
	  ] .

Though the rdfa:defines, rdfa:ProfileDocument stuff would not affect
how terms and prefixes are extracted from a profile, so would be
effectively optional stuff, just added by people who want their
profiles to have added semantic goodness.

Also, I think ideally the range of rdfa:uri should be xsd:string, not
xsd:anyURI. This is because CURIE prefix expansions could theoretically
be too truncated to be legal URIs until the suffix is added.

Toby A Inkster

Received on Tuesday, 5 October 2010 16:06:49 UTC