Re: Making querying of annotations optional

> On 15. Aug 2024, at 17:56, Olaf Hartig <olaf.hartig@liu.se> wrote:
> 
> On Thu, 2024-08-15 at 15:41 +0000, Franconi Enrico wrote:
>>> On 15 Aug 2024, at 17:25, Thomas Lörtsch <tl@rat.io> wrote:
>>> 
>>> Am 15. August 2024 16:45:29 MESZ schrieb Franconi Enrico <
>>> franconi@inf.unibz.it>:
>>>> Before letting this discussion go too far, I want to be sure that
>>>> we share the same assumptions.
>>>> [0] assumes a CG-style notion of triple reification, which is not
>>>> the one adopted by the current baseline.
>>> 
>>> I don't think so: IIUC embedded triples in [0] are asserted and
>>> referentially transparent types, wheras in the CG report they are
>>> unasserted and referentially opaque types.
>> 
>> [0] does not define a semantics for RDF*, nor simple entailment, but
>> syntactically [0] has an abstract syntax mirroring the CG syntax,
>> where there is no distinction between triple terms and triple
>> reifiers; and SPARQL* in [0] is not based on BGP matching.
> 
> SPARQL* in [0] is based on BGP matching, with the addition that it also
> takes into account the triples that are contained within other triples
> of the queried graph.
> 
>> I’ve nothing against [0], I am only observing that is quite far away
>> from our baseline.
> 
> Yes, and you can happily ignore this branch of the email thread which I
> only opened to quickly react to something that Thomas wrote about [0].
> It is not really relevant to the current discussions of our group.

IMO it is because
- it shows that querying triple and triple terms together has been deemed feasible already in RDF*/SPARQL* implementations
- the fact that this poposal got so popular is a use case description in its own right
.t



> Best,
> Olaf
> 
>

Received on Thursday, 15 August 2024 16:03:03 UTC