Re: what about RDF format?

On Wed, Mar 23, 2005 at 10:48:03AM +0000, Steve Harris wrote:
> 
> On Wed, Mar 23, 2005 at 09:49:52 +0000, Andy Seaborne wrote:
> > Yoshio,
> > 
> > For testing, we have been using
> > 
> > http://www.w3.org/2001/sw/DataAccess/tests/result-set.n3
> > 
> > because tests evaluate to a graph for SELECT and CONSTRUCT,
> > then one way to check the results is to do a graph comparison.
> > 
> > THis vocabulary can't cope with sorted results.
> > 
> > This could form the basis of an RDF form.
> 
> It certainly makes sense that they be the same. The order problem has been
> bothering me, as the testcase result format will need to be able to repesent
> this if we get ORDER BY. I would prefer the binding results to not be in a
> rdf list due to the difficulty of processing the graph, and sequences seem
> to be unfashionable.
> 
> Are structures like:
> 
>         [] :list [ :position 1 ;
>                    :value    <http://foo.example.com> ];
>                  [ :position 2 ;
>                    :value    <http://bar.example.com> ]
> 
> equally unpopular? Its my current preferred fix. The schema and results
> files can be modified quite simply to represent order in this way, and its
> easy to query.

Easy to query provides a way for RDF tools that don't have SPARQL
interfaces to use a native API or query lang to get at the results.
For instance, TAP could GET a specific resource like
  http://sparql.org/?q="..."
and use it's API to get at the results.


However, since the results are XSLT-able, we can use some emergent
XSLT to transform results format into RDF. Thus, the above example
query would turn into
  http://xslt-service.org/?template=SPRes2RDF&doc=http://sparql.org/?q="..."

Thus, I propose we do nothing about the RDF format because it can be
done post-standarization by interested parties.

> The other thing that the result format cant represent is the order of the
> columns, which is probably less of an issue, but rq23 does say something
> about it, as the XML result format can distinguish.

XSLT can count. I've seen Dom make it count, but Dom gets it to do
lots of things that mortals can't. Anyways, I think your posN solution
would work for the order of columns as well.
-- 
-eric

office: +81.466.49.1170 W3C, Keio Research Institute at SFC,
                        Shonan Fujisawa Campus, Keio University,
                        5322 Endo, Fujisawa, Kanagawa 252-8520
                        JAPAN
        +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA
cell:   +81.90.6533.3882

(eric@w3.org)
Feel free to forward this message to any list for any purpose other than
email address distribution.