Re: evaluating SPARQL w.r.t an RDF query language survey

On Thu, 2005-04-07 at 12:46 +0200, Jeen Broekstra wrote:
> On Apr 6, 2005, at 12:07 AM, Dan Connolly wrote:
> > A comment asks...
> > "How many of the 14 test questions in [1] does the current SPARQL spec
> > cover ?"
> >  --
> > http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2005Apr/0005.html
> >  -> http://www.aifb.uni-karlsruhe.de/WBS/pha/rdf-query/
> > 
> > I'm interested to take a crack at it, but I'd like some confirmation
> > from other WG members that I've got this right before I reply.
> > Perhaps it's worth making these into test cases...
> 
> I am one of the co-authors of that report, so if I can help out in any 
> way let me know.

I think it would be great for you to add these cases to our test
suite.
 http://www.w3.org/2001/sw/DataAccess/tests/

If you're interested, send an ssh2 public key to EricP and me
and we'll try to arrange CVS write access. I presume you can
get in touch with Steve to learn anything you don't yet know
about how our test suite works.

> [snip]
> 
> > 5 Quantification
> > Return the persons who are authors of all publications.
> > 
> > nope.
> > 
> > The langauges with "yes" in this column seem to
> > support closed-world assumptions that
> > we don't support in SPARQL.
> 
> I'm guessing that you are referring to the FORALL operation used in 
> the example RQL query? Or more generally, to the notion of "all 
> publications" in the specification of the query?

Well, yes. The query says "all publications" but the "yes"
answers only regard "all publications mentioned in the file".

> Is universal quantification by definition something that falls outside 
>   SPARQL,

No; I'm not aware of anything in our charter that makes it
necessarily out of scope; it's just that we haven't had
use cases, requirements, objectives, or design proposals
in that direction.

>  or can it be scoped (for example, employing quantification 
> on a named graph to allow closed-world assumption on that particular 
> graph)?

I'm not sure.
Maybe there's some way to do the "all publications mentioned
in this file" query with bound() and OPTIONAL too. I don't
have a good feel for the expressive capability of those two yet.

> > 6 Aggregation
> > Count the number of authors of a publication.
> > 
> > nope.
> > 
> > unique names assumption.
> 
> I have trouble with this. I think that defining a counting operation 
> that does not count entities, but simply labels (URIs, bNodes, 
> literals), would be useful. This does not make any unique names 
> assumption, AFAICS. Counting would simply be a way of retrieving the 
> number of results a query would give, without giving the actual result.

Well, perhaps, but that's not how I understood the test query.
It says "count the number of authors," not "count the number of
terms used in the given data to refer to authors".

> Jeen
-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Thursday, 7 April 2005 13:49:51 UTC