Re: RDF-star and Conjectures

Dear all, 

Last message, I promise. 

> On 17 Sep 2021, at 11:54, Andy Seaborne <andy@apache.org> wrote:
> 
> 
> 
> On 16/09/2021 18:07, Fabio Vitali wrote:
>> Dear all,
>> My main reason of interest in RDF* is the mechanism to express a triple without stating it (quoted triples, formerly embedded triples). I think that that postulating (i.e., expressing without asserting) is an extremely worthy thing to have and work with. I think that it is truly a fundamental and innovative aspect of RDF*.
>> Still, I am concerned that this applies only to individual triples and that annotations are considered just a syntactic trick to represent at the same time both a quoted triple and its assertion.
> 
> It is only syntax, why does that make it a "trick"?
> 
> rdf:type is syntax for <http://www.w3.org/1999/02/22-rdf-syntax-ns#type>
> 
> 123 is syntax for "123"^^xsd:integer

Because at the moment annotations are not referenceable. That is not a problem, until someone starts wondering how to express indirect claims: "Fabio Vitali asserts that Samuel Johnson's claim is true". 

You are sponsoring that quoted triples are provided with an URI, so as to be able to make references to their specific occurrence. Well, I am wondering if annotations deserve to have their own URI as well *as an annotation* and not as a referenceable quoted triples + a non referenceable plain triple. 

>> I am wondering, specifically, if the fact that quoted triples are postulated
> 
> Here, postulated is presuming about the claim.
> 
> Sometimes, we really do want so refer to the exact form of the triple - when on the web, the data may have different entailment regimes and assumptions.
> 
> Only if everything is local can you switch to the logical level only.
> 
>> but not asserted came out as a pleasant side effect of the peculiar type of reification adopted for the formal model of RDF* triples, or if this feature was (is, will be) considered as a true design goal for the group, and if there are plan to explore this specific subtree of additional features.
>> If this is the case, then I am wondering if there is an interest in discussing sooner or later about the following:
>> 1) promote annotations to first class citizens of the model, and not just leaving them as syntactic shortcuts
> 
> We loose the possibility of referring to a triple without assertion which arises in provenance. On the web, your inference assumptions != my inference assumptions.

This is exactly what I am trying to say. But give an annotation a URI, and we can distinguish between giving provenance to an assertion and to an annotation containing that assertion. I think this would be useful. 

> Conjectures are more akin to RDF Reification where a reification is referring to the logical assertion whereas RDF-star is quoting the form.

No no and no. Conjectures by design do not use any form of reification. They use secondary predicates.  

> In fact, I think access to the quoted form is needed for conjectures if it is to be a web technology.

Totally agree. 

> Either a system which has graphs as RDF terms (N3) or the ability to refer to part to the contents of a conjecture is necessary on the web where there is data published by other people and not under the users control. How to describe and refer to elements of that other published data?

Conjectures assume that a graph is designed as being conjectural, and therefore can be asserted (collapsed) or negated all at once with a single step. Being graphs, they are by definition addressable. The same is not true for N3 graphs, which are not individually addressable, but can only be placed inline as subjects or objects of other triples. 

So with N3 it becomes rather complicated to negate someone else's graph. I could, in theory, repeat your graph and negate it, but that would be a different resource, i.e., a different graph, not a negation of your graph. 

Thanks

Ciao

--

> 
>> 2) extend the reach from postulated triples to postulated graphs
>> 3) include mechanisms to postulate the postulation of triples/graphs (already doable)
>> 4) include mechanisms to postulate the assertion of postulated triples/graphs
>> 5) work with the cascade effect of asserting chained postulations
>> If there is an interest in some of these topics, I would very much like to take part and contribute actively to the discussions, and I would like to know how and where and when should the discussion items be introduced.
>> Thank you
>> Fabio Vitali
>

Received on Friday, 17 September 2021 14:33:35 UTC