RE: another update test added (was: RE: Questions on grammar restrictions on Blank Node reuse across...)

Hia again,

> Does not change anything.  It does not create a shared bNode.

I don't want to test shared bnodes, because - as I think you agree -  this is not
expressible. I want to approximate this (just as 05a tries to approximate this).
I think that insert-05 is a closer approximation than insert-05a, that's why
I prefer to have 05 in.

Hope that clarifies matters,
Axel


> -----Original Message-----
> From: Andy Seaborne [mailto:andy.seaborne@epimorphics.com]
> Sent: Tuesday, 10 July 2012 3:11 PM
> To: Polleres, Axel
> Cc: public-rdf-dawg@w3.org
> Subject: Re: another update test added (was: RE: Questions on
> grammar restrictions on Blank Node reuse across...)
>
>
>
> On 10/07/12 13:59, Polleres, Axel wrote:
> > Hi Andy,
> >
> > I don't want to hold anythnig up here, would the following
> fix within
> >
> http://www.w3.org/2009/sparql/docs/tests/data-sparql11/basic-update/in
> > sert-05a-g1-pre.ttl change your negative perception of insert-05?
> >
> >      s/_:b1/[]/
>
> Does not change anything.  It does not create a shared bNode.

I don't want to test shared bnodes, because - as you say, or no? - this is
not expressible. I want to approximate this.

> >
> > That would make clear that we are not talking about shared
> blank node labels here.
> > Would that be a compromise?
>
> No - we are talking about the same blank node.  That's the
> point of the test!

The point is to approximate this, and I think 05 is a closer approximation than 05a.

.
> >
> >
> >> " in fact, does not actually test the right thing"
> >
> > I am frankly not sure what you mean by "the right thing"
>
> shared blank node.
>
> > Just to recap, what I was aiming at testing here is whether
> the double
> > insert is idempotent.
>
> and results in a shared blank node.
>
> > 05 tests this, i.e. it tests whether the resulting graphs stay
> > equivalent according to the definition
> (http://www.w3.org/TR/rdf-concepts/#section-graph-equality).
> >
> > 05a does *not* test this, is tests whether the resulting
> graphs have
> > the same number of triples, not whether they are
> equivalent. I do not
> > oppose test case 05a at all, it is fine to approve it, but
> it tests a weaker condition.
>
> It avoids needing to define dataset isomorphics for this one test.
>
> Under dataset isomorphism, 05 is wrong and should be failed.
>
> The test README weakens testing to make it easier to
> implement.  We should not define SPARQL by the weaken testing.
>
> > That's why I'd prefer to have 05 in as well (with the
> proposed change above, if that's something people can live with).
>
> Why?  What does it add other than a test that should be failed!
>
> Do you have a system that requires 05 effects?
>
>       Andy
>

Received on Tuesday, 10 July 2012 13:20:02 UTC