Re: Update

good evening;

> On 2015-10-12, at 18:54, Gregg Kellogg <gregg@greggkellogg.net> wrote:
> 
>> On Oct 11, 2015, at 2:00 PM, Karima Rafes <karima.rafes@gmail.com> wrote:
>> 
>> […]
> 
> I’m not sure why a proprietary service is necessary to describe and run such tests;

the particular test feature at issue at the start of this thread was “federation”.
there are two senses in which the process which is implicit in the “run the manifest” paradigm is markedly deficient.

the manifests, as written, stipulate a location which does is not a sparql service.
if we take the example, service01.rq, we see

>     # SERVICE join with pattern in the default graph
> 
>     PREFIX : <http://example.org/> 
> 
>     SELECT ?s ?o1 ?o2 
>     {
>       ?s ?p1 ?o1 .
>       SERVICE <http://example.org/sparql> {
>         ?s ?p2 ?o2
>       }
>     } 
> 

from which, one might expect a request on the order of

    curl -H "Accept: application/n-quads" http://example.org/sparql

to yield a service description.
it does not. it yields a narrative encoded as html.

if nothing else, if the service directory remains, there should be a host which serves the specified location(s) - even if as static files.

the second sense in which the current process falls short is that, if the intent is to facilitate interoperability by demonstrating it and by isolating cases where it is not present, then there must be a way to test combinations of purportedly conforming, live implementations.
for this purpose, a service on the order of “sparqlscore” would add value to the test suite, whatever its aspirations may be and whether or not it is proprietary.

> […]

best regards, from berlin,
---
james anderson | james@dydra.com | http://dydra.com

Received on Monday, 12 October 2015 18:04:20 UTC