Re: followup to today's meeting

On 20/04/2017 11:01, Sandro Hawke wrote:
> The answer is that if we're able to delineate a safe subset of SPARQL 
> for use in SPARQL-based Constraints, that would not require restarting 
> the CR clock.

That's great news!

>    I'll be very interested to see if we end up with consensus that 
> that's technically feasible (eg that Peter agrees the subset is safe).

I'll work on this with Andy, who is unfortunately on holiday for a bit.

>
> Meanwhile, the test results [1] for core/* look amazing.    Very close 
> to CR-exit for Core right now.   It would be good to understand why we 
> have no-data on two, and what's going on with the failed Netage tests, 
> but still it looks good.

The tests have simply not been executed by everyone yet.

>
> The test results for sparql/* are more modest.   Do we know the 
> stories here?  Holger, are the failing TopBraid tests ones you don't 
> think should be in the test suite for some reason?

Yes, all SPARQL tests that currently fail (and some more) would become 
invalid because their constructs (e.g. this scenario with MINUS and 
possibly the nested SELECT) would be made undefined. But I still need to 
discuss this idea with Andy.

>   ("shared" we discussed; pre-binding/minus sounds like one you want 
> to exclude from the language; not sure about pre-binding/bind.)   Do 
> other folks plan to implement the sparql stuff?   If we're not sure, 
> that's another reason to try to separate sections 5 & 6 into a 
> different document.
>
> Where are we on positive and negative syntax tests being in the suite 
> & reporting?

I see no need to formalize them, and would be strongly against making 
this yet another thing on our plate. If an engine is capable of 
executing SHACL then it will also be able to validate the shacl-shacl 
graph, so there is IMHO no need for everyone to repeat that same 
tautological test. They just need to verify SHACL Core compliance.

Holger


>
>    -- Sandro
>
> [1] 
> http://w3c.github.io/data-shapes/data-shapes-test-suite/#validate-rdf-data-tests
>

Received on Thursday, 20 April 2017 01:13:11 UTC