Re: on entailments and triple terms

Just a thought, but wouldn’t it be more convenient if triple terms were treated as being asserted unless they explicitly indicate they belong to a different graph?  This assumes that the common use case is to annotate information where the annotations are held to be true.  If you want to model counterfactuals it would be reasonable to introduce named graphs.

> On 19 Aug 2024, at 22:23, Thomas Lörtsch <tl@rat.io> wrote:
> 
> 
> 
> Am 19. August 2024 12:52:30 MESZ schrieb Franconi Enrico <franconi@inf.unibz.it <mailto:franconi@inf.unibz.it>>:
>> On 19 Aug 2024, at 12:42, Thomas Lörtsch <tl@rat.io> wrote:
>> On 19. Aug 2024, at 12:26, Franconi Enrico <franconi@inf.unibz.it> wrote:
>> On 19 Aug 2024, at 12:16, Thomas Lörtsch <tl@rat.io> wrote:
>> 
>> RDF* and SPARQL* are defined to require something that might be called a simple * -entailment: any RDF triple is also an RDF* triple [0]. That captures a plausible and useful intuition. It has been implemented. It was successfull. It was deemed useful for bridging the RDF/LPG divide. This WG was chartered to standardize the approach.
>> 
>> As it has been noted also by Olaf, [0] does not relate at all to the current baseline.
>> Moreover, in the current baseline triple terms are NOT triples.
>> Simple *-entailment is not simple entailment as defined in the current baseline.
>> SPARQL should be based on simple entailment (unless there is an explicit entailment regime).
>> Simple entailment in the current baseline is just BGP matching.
>> 
>> I’m clearly not arguing that the current baseline is good and ready, but that things need to be changed and extended. And you said on Friday that we shouldn’t use reification (which the current baseline is based on). So what exactly are you suggesting?
>> 
>> I am arguing that your proposal to have SPARQL-* is not acceptable if we keep the current baseline.
> 
> Yes, you said that already. And again I'm replying: then let's not keep but change the current baseline. 
> 
>> The current baseline is compatible with the proposal to extend RDFS with rdfs:states as in [1]
> 
> Yes, but that means that what we need is only available in RDFS. I think we can say that indeed to correctly - by the book - express statements about statements in RDF we need entailment. But we can also say that on the lower level of simple RDF we can emulate RDFS entailment through annotation syntax and a SPARQL* inspired approach to querying. 
> 
> Or we just give up, say reification is all we have (and you shouldn't use it). In that case it would be better to dissolve this WG and keep the syntactic options unencumbered by an insufficient approach. Yes, I mean that. Reification is not the right tool. Any attempt to use it in earnest for what people will want and expect to use it - annotating asserted statements - will be easy prey to semanticists that claim that the reification refers to whatever but not the triple in the graph. That, frankly, is fucked up. We already have a metamodelling approach with nice syntax but no clear semantics: named graphs. We sure as hell don't need another failure like that. 
> 
> If you and others in this group see E/R style modelling as the way out of this problem then you're free to pursue that approach. But that is no excuse for messing up what this WG is tasked to achieve: statements about statements, and bridging the gap between RDF and LPG. 
> 
> . t
> 
> 
> 
>> —e.
>> 
>> [1] https://github.com/w3c/rdf-star-wg/wiki/Extending-the-baseline-with-%22asserted%22-stuff
>> 
>> [0] Olaf Hartig: Foundations of RDF* and SPARQL* - An Alternative Approach to
>>  Statement-Level Metadata in RDF. In Proceedings of the 11th Alberto Mendelzon
>>  International Workshop on Foundations of Data Management (AMW), Montevideo,
>>  Uruguay, June 2017
>>  http://olafhartig.de/files/Hartig_AMW2017_RDFStar.pdf

Dave Raggett <dsr@w3.org>

Received on Tuesday, 20 August 2024 10:47:59 UTC