Re: Annotation syntax

On 17/03/2024 13:07, John Walker wrote:
> Hi Andy,
> 
> Thanks for the response.

Firstly - none of this is finally decided - it is work-in-progress.

Annotation syntax recognizes that asserting and making statements about 
the usage of a triple will be common.

The idea of having annotation syntax has been reasonable stable.

However, not allowing making statements without asserting does not cover 
enough of the possibilities.

> It seems I need to relearn the semantics of the RDF/SPARQL-star syntax 
> from the earlier WG.
> 
> Are any constraints applied to where the <<( :s :p :o )>> triple concept 
> terms may be used.
> 
> Is that only in the “object” position like literals, or also as a 
> “subject” like IRIs and blank nodes?

Currently, object position, because it is a "literal like", but this is 
not decided yet. There are arguments both ways.

> Or even as a predicate?

The predicate position is always an IRI; it's fundamental to RDF and it 
is part of how the semantics of RDF work. It's the set IP in

https://www.w3.org/TR/rdf11-mt/#simple-interpretations

> Given generalized RDF triples, graphs, and datasets [G-RDF].

Undecided.

And generalized RDF is non-normative.

> Is there anything to be said to only allow triple concept terms as an 
> object in non-generalized RDF?
> 
> Further still, why not just use RDF literals for triple concepts and 
> introduce a new datatype (cf. rdf:HTML, rdf:XMLLiteral, and rdf:JSON)?

That has been considered.

> Apologies if I retread old ground, it’s hard to keep track of what has 
> and has not been discussed based on the mailing list.

The working group has live working drafts of documents. Updates for the 
current direction of the working group are in-progress. There are a lot 
of documents ...

> [G-RDF] https://www.w3.org/TR/rdf12-concepts/#section-generalized-rdf 
> <https://www.w3.org/TR/rdf12-concepts/#section-generalized-rdf>
> 
> Regards,
> 
> John Walker

Received on Sunday, 17 March 2024 18:26:06 UTC