Re: Test without data?

Dear Axel,

On 11/22/2012 08:18 PM, Polleres, Axel wrote:
> Dear Jan, Jeen,
>
> this is a joint response to your comments
>
> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2012Oct/0022.html
>
>
http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2012Oct/0026.html
> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2012Oct/0028.html
>
>  We have added the following clarifying note to the README file
> describing Query Evaluation tests, cf.:
> http://www.w3.org/2009/sparql/docs/tests/README.html#queryevaltests
>
> "In case the query in the qt:query predicate contains FROM and FROM
> NAMED clauses and no qt:data is present, the graphs comprising the
> test's RDF dataset are expected to be loaded by dereferencing the
> respective URIs of the FROM/FROM NAMED clauses."
>
> We'd appreciate if you could let us know in a brief response whether
> this addresses your concern.

Well, it clarifies the situation.  But, I fail to see the logic why
the test doesn't simply include a qt:data statement.  Some of the tests
do (AFAIK).  Why not all?  Now the test driver cannot simply prepare
the data.  FROM, neither FROM NAMED defines that it should load any
data.  You force this unspecified behaviour on the test framework, asking
the query evaluator to execute it or the test driver to parse the
queries and perform the loading.

For short, I think that adding a couple of qt:data statements is a
much simpler and better solution.  It simplifies writing the test
driver and most likely saves a lot of time for testers to deal with
this issue.

	Regards --- Jan

>
> Axel, on behalf of the SPARQL WG
>

Received on Thursday, 22 November 2012 19:37:50 UTC