– DRAFT –
RDF-star Semantics TF

Discussion on weaker notion of opacity: IRI-opaque triple terms

Introduction to baseline with transparent and IRI-opaque triple terms, and no functionality of annotations

<pchampin> we could avoid the complexity of TEPs in general, and decide that rdf:reifies is the only TEP

<AndyS> pchampin - Interesting

<AndyS> The baseline abstract grammar (as of 15:17 today) has one syntactic triple term (i.e. writable). The "transparentTripleTerm | opaqueTripleTerm" has no visible effect. They can be distinguished in well-formed (by usage). And we are at pchampin suggestion.

<pchampin> +1 to niklas

<pchampin> two kinds of provenance

<pchampin> AndyS for me the unrestricted abstract syntax implies that concrete syntaxes should be able to distinghishvthe two kinds of triples

<pchampin> if rdf:reifies is many-to-many and rdf:annotationof is one-to-many, I don't think that this property chain works

<niklasl> +1 to souri (same point as Gregg)

<pchampin> n-triples !

<pchampin> and turtle can use raw triple terms

<pchampin> that's not forbidden

<niklasl> exactly

<Souri> Never bet on what users will NOT do! :-)

<niklasl> +1 to 100% of what pchampin said

<pchampin> +1 Souri

<niklasl> +1 to Souri too

<pchampin> bye

<Souri> If we go with two distinct properties (instead of going with two types of tripleTerms) -- rdf:annotationOf and rdf:reifies -- can we have the same tripleTerm used with two reifiers? => :r1 rdf:annotationOf <<( :bob :likes :alice }>> . :r2 rdf:reifies <<( :bob :likes :alice )>> . We can is my guess.

<tl> @tallted will do, thanks for your agreement!

<tl> https://link.springer.com/chapter/10.1007/978-3-031-47240-4_15

<tl> Iglesias 2023 Comparison of Knowledge Graph Representations for Consumer Scenarios

<Souri> :r1 rdf:annotationOf <<| :bob :knows :alice |>> .

<AndyS> dstl/IES4

<niklasl> :r1 rdf:subject <bob-at-21> . <bob-at-21> prov:specializationOf :bob .

<Souri> :r1.subject :age 21 . :r1.object :age 18 .

<niklasl> .. terribly, decicdedly exdurant.

<niklasl> Btw, age is generally a very bad predicate (it locks the subject interpretation to an entity in a moment of time).

<Souri> :r1=>subject :age 21 . :r1=>object :age 18 .

<niklasl> Different interpretations of :honoraryTitle <resource> . and that resource used as a relationship to a country....