Re: primer review part 2 (read me first)

On 6/11/07, Jeremy Carroll <jjc@hpl.hp.com> wrote:
> Brian Suda wrote:
> >    <Description xmlns="http://www.w3.org/1999/02/22-rdf-syntax-ns#" rdf:about="_666">
>
> Question:
>    Do you really want to be creating a new URI here?

--- i think we had a debate about this, i can't seem to find the
reference now. The issue was one of a few things...

1) use the ID attribute, but this is not required for the hreview, so
not every instance will have an ID to act as a hook.

<div class="hreview" id="_666">...</div>

If there was an ID then the rdf:about should be:
http://www.w3.org/2001/sw/grddl-wg/doc29/hotel-data.html#_666

2) if there was not an ID we considered making it a bnode, or use the
URL that might be supplied in the review itself. If you are reviewing
http://example.org/marriott then it makes sense that it is URI of what
you are describing in the rdf:about.

3) what if there is no ID and no URL, then it could be a bnode, but
then i'm not sure how that effects collapsing multiple reviews of the
same place. If i review Hotel-A with 3 stars, and you review it with 5
stars, we need to make an average of 4. This is not used in the SPARQL
in this example, but the XSLT should generate RDF that this COULD be
done - but maybe this isn't important for the primer.

> Does the Primer include SPARQL results and example SPARQL queries
> against this GRDDL result. If so, those queries and results may need
> updating.

--- once we choose how to update the output, if the SPARQL is not
working, i can correct it. I wrote the original queries, so if we need
changes i can certainly jump in.

-brian

-- 
brian suda
http://suda.co.uk