– DRAFT –
RDF-star Semantics Task Force

<TallTed> scribe:

Starting the work to come up with a concise proposal to eventually submit to the WG

<AndyS> The W3C calendar does not always show TF meetings. "Working Group - Calendar - include TF apply" does not show TF. My personal calendar does.

<Zakim> AndyS, you wanted to ask what we are trying achieve today

ex: spouse1, ex:spouse2 replacing ex:spouse would require remodeling and/or rewritten tools

<niklasl> Thanks Ted; yes, of course.

<Souri> +1 to AndyS sharing a notepad or something showing outline as it evolves

<Souri> -1 to breaking Turtle is an extension of N-Triple

<Souri> +1 to need for new patterns in SPARQL -- I called it "edge-pattern", in addition to "triple-patterns" (going with "Triples" vs. "Edges") -- ?s ?p ?o | ?n .

the two goals as we discussed previously — annotation (talking about something asserted) and quotation (talking about something _not_ asserted) — see near the end of https://www.w3.org/2023/11/16-rdf-star-minutes.html

<tl> 2 words: type and instance

in other words, "ROWID is automatic"

saying "Liz marriedTo Richard" 17 times doesn't deliver what's needed

<tl> tallted x likes y 17 times might make a differece to someone

<niklasl> For 1:1 LGP "import": <Liz> :marriedTo <Richard> {| |}, <Richard> {| |}, ... . # ... 15 more times

tl -- I think that's where named graphs come in. `x likes y` can only occur once in a given graph, but it could occur in 17 graphs (e.g., one per "speaker").

<niklasl> Ted: yes, and therein lies some kind of "isomorphism" between "occurrences" and named graphs (the difference lies in assertedness, and in consequence what entailment acts upon)

<tl> tallted nested graphs use named graphs that way, but i just wanted to make the point that i may be of interest, although it's all about the same type

<tl> niklasl - i don't disagree :)

<niklasl> +1 you need to say *something* more; even if it is just 17 different blank names for occurrences of such a triple

<Zakim> enrico, you wanted to show an example and discuss about it - to make the discussion more concrete

<niklasl> +1 for example-oriented

<enrico> :liz :spouse :richard .

<enrico> << :wed-1 | :liz :spouse :richard >> :starts 1964 ; :ends 1974 .

<enrico> << :wed-2 | :liz :spouse :richard >> :starts 1975 ; :ends 1976 .

<enrico> << :wed-1 | :richard :spouse :liz >> .

<enrico> << :wed-1 | :liz a :wife >> .

<enrico> << :wed-1 | :richard :spouse :liz >> ; :starts 1964 .

<<: wed-1 | :liz :spouse :richard >> :starts 1964 ; :ends 1974 ; :weddingLocation :Montreal .

<<: wed-2 | :liz :spouse :richard >> :starts 1975 ; :ends 1976 ; :weddingLocation :Botswana .

<enrico> << :wed-1 | :liz a :wife >> :starts 1964 .

<enrico> << :wed-1 | :richard :spouse :liz >> :starts 1964 .

<Souri> :wed-1 | :liz a :wife . :wed-1 | :richard :spouse :liz . :wed1 :starts 1964 . :wec-1 :ends 1974 .

<tl> talllted: maybe you want a reference to the type (without asserting it) that would be used eg to aggregate the number of occurrences (without creating a new occurrence itself)

<ora> I have to drop.

+1 olaf

<AndyS> +1 to Olaf

<niklasl> :wed-1 rdfx:occurrenceOf << :liz a :Wife >>, << :richard :spouse :liz >> .

<AndyS> <<( :liz a :Wife )>>

<tl> andySi think that syntax is too involved

<niklasl> :wed-1 rdfx:occurrenceOf <<( :liz a :Wife )>>, <<( :richard :spouse :liz )>> .

<Souri> We could focus on this doc on the mailing list.

<niklasl> It is defined like that for an RDF merge, but not in the graph protocol ("PUT data as graph")

<niklasl> In the second case it is "overwritten"

<Souri> Even FROM :G1 FROM :G2 would merge two named graphs into a default graph (in SPARQL).

+1 niklasl

<AndyS> PUT -> POST

<niklasl> Ah, yes. Also, defined as merge for an RDF (dataset) source (a merge is a merge :) )

<niklasl> Just to note what I said previously about the suggested "named annotations" syntax: it collides with this use of AlternativePath in an annotation query: SELECT * { ?s ?p ?o {| dct:issued | dct:modified "2023" |} . } # <- that is not a name, its either dct:issued or dct:modified

<AndyS> https://github.com/w3c/rdf-star-wg/wiki/Triple%E2%80%90Edge-subgroup-proposals

<gkellogg> https://github.com/w3c/rdf-star-wg/wiki/SubGroup-Proposal

<niklasl> Be able to talk about something asserted, and something unasserted, alike.

<niklasl> (something(s))

souri wanted to ask whether we agree: occurrence (or edge) s-p-o-n (with annotations) does NOT necessarily imply presence of (=> asserted) s-p-o triple.

<Zakim> Souri, you wanted to ask whether we agree: occurrence (or edge) s-p-o-n (with annotations) does NOT necessarily imply presence of (=> asserted) s-p-o triple.

<niklasl> Cf. GRAPH ?x { ?s ?p ?o } to know if ?x is a name of a graph.