W3C

RDF-star

17 September 2021

Attendees

Present
AndyS, Fabio_Vitali, gatemezing, olaf, ora, pchampin, rivettp, TallTed, thomas
Regrets
james
Chair
pchampin
Scribe
ora, pchampin

Meeting minutes

<AndyS> We must be approaching 5 sigma on this by now.

<thomas> if Ora in room -> scribe Ora

Announcements and newcomers

<pchampin> https://mosaicrown.github.io/scg2021/

pchampin: Last week's workshop...

pchampin: It went well, lively discussion, interesting things, papers will be on-line soon.

pchampin: Presentations were recorded.

pchampin: It was a hybrid conference, it was great to see people.

pchampin: There is a lot of interest for RDF-star.

olaf: Announcing that I have been selected to be an Amazon Scholar, I will work with the Amazon Neptune team. Still associate professor in Linköping too.

olaf: RDF-star is one of the topics.

Open actions

<pchampin> https://github.com/w3c/rdf-star/issues?q=is%3Aopen+is%3Aissue+label%3Aaction

<pchampin> https://github.com/w3c/rdf-star/issues/195

pchampin: Confirmed that process does not limit what a new feature can be. It can be limited by the charter.

<pchampin> https://github.com/w3c/rdf-star/issues/196

ora: proposed a text explaining the distinction btw asserted and quoted
… made a parallel with evaluated/quoted in LISP

AndyS: agreeing with what ora is saying
… going back to the spec: a triple is a 3-tuple, that's all
… there is an infinite set of it, "floating around"
… they are asserted once put into a graph
… in a given graph, a triple maybe both asserted and quoted

pchampin: I do not agree that the chevrons are a quoting operator, they are an embedding operator.

pchampin: In Lisp, you can simply remove the quote. Not so in RDF-star.

pchampin: Context of use makes a triple quoted or asserted.

pchampin: I do not disagree with the analogy, but it could be confusing.

<TallTed> Nitpick: Chevrons are the `«»` glyphs, which are distinct from double angle brackets `<< >>`.

<AndyS> "A quoted triple is a triple used as the subject or object of another triple."

<Zakim> AndyS, you wanted to bring in some current spec text

AndyS: This line from the spec... change it to "...can be used...".

AndyS: Chevrons are doing the quoting.

thomas: Aren't we making things too complicated?

pchampin: I am not making a distinction between quoted and embedded triples.

pchampin: The main role of the chevrons are to "embed" a triple. You can only do this in a certain context.

thomas: Also you can only quote in a certain context.

pchampin: Reluctant to see chevrons as a quoting operator.

<AndyS> Writing a triple as (:s :p :o) then ( (:s :p :o) :q "abc" )

AndyS: Everyone has a different background, everyone's viewpoint is different.

AndyS: Is there anything wrong in the spec?

AndyS: What do we have to do, to change in the spec?

pchampin: Thinking was that something might be missing.

pchampin: I am not against adding Ora's paragraph.

pchampin: Details can confuse...

pchampin: If we think this adds to folks' understanding, then why not.

pchampin: Defer the decision as not everyone has read it.

<Fabio_Vitali> I can't understand where is the text going to end up...

pchampin: Ora please make this into a pull request.

pchampin: Mark as non-normative.

ora: OK

AndyS: I particularly like the last sentence.

Action: ora to turn his proposal in #196 into a PR

<pchampin> https://github.com/w3c/rdf-star/issues/201

pchampin: I did this.

pchampin: Nice side effect was creating some publicity for the Easier RDF repo, https://github.com/w3c/EasierRDF

pchampin: My feeling is the consensus is "why not, but let's be careful".

<AndyS> +1

pchampin: Next two actions are mine...

pchampin: Not everyone has reviewed the PRs.

pchampin: Still have to push the other one.

PR #204: Classes for RDF-star terminology

<pchampin> https://github.com/w3c/rdf-star/pull/204

pchampin: Fleshing out the vocabulary section.

AndyS: I took existing material, refined a bit.

AndyS: Discuss classes first, before discussing properties.

thomas: What is the source?

pchampin: <scribe missed the answer...>

thomas: Is that another term for graph?

pchampin: No, because of mutability.

thomas: I read it but meaning was not obvious.

pchampin: This is still bare, just trying to define terms not yet defined (RDF-star source).

thomas: It is difficult to discuss terms that are a mathematical abstraction but also very concrete.

thomas: No sense discussing this yet.

<pchampin> PROPOSED: merge #204

<pchampin> +1

<olaf> +1

<AndyS> +1

<ora> +1

<gatemezing> +1

<TallTed> +1

<thomas> -1

<rivettp> +1

<Fabio_Vitali> My list of issues goes up to #203. Where do I find 204?

<gatemezing> AndyS: I guess we don't need to add axioms like disjointness between some clases?

<Fabio_Vitali> anyway +1

Resolution: merge #204

Defining our path

pchampin: From AndyS' email: we should give ourselves some milestones, we are soon one year old, we have achieved a lot.

pchampin: Also the next step: chartering a WG.

pchampin: If all goes well, a future WG will take this up.

AndyS: I predict the WG will revisit a lot of things.

AndyS: Limited value in getting into really fine details.

AndyS: REC-grade doc not needed now.

AndyS: Early on in the WG there will be discussion by new people about what RDF-star really is.

<TallTed> and what RDF-star is NOW, vs original and other earlier papers...

AndyS: We cannot possibly have considered everything.

thomas: Issue #170, referential opacity, is crucial.

thomas: Makes a difference in many use cases, such as provenance.

pchampin: A solution or a discussion?

thomas: A proposal, a solution.

<AndyS> 170

thomas: We should evaluate how much effort this is.

AndyS: Another source of validation exists.

thomas: The proposed semantics does not cover usage.

pchampin: Evidence?

thomas: Look on the Web.

thomas: I do not have to provide evidence.

thomas: Want a proof of how referential opacity is easy.

thomas: Something that is worked out.

<olaf> https://github.com/w3c/rdf-star/issues/202 is still an open action on me

<pchampin> https://github.com/w3c/rdf-star/issues/202

thomas: I need an answer, then I can comment.

<pchampin> https://w3c.github.io/rdf-star/cg-spec/2021-07-01.html#alternatives-to-referential-opacity

pchampin: I did some prototyping in N3. I understand it is not a standard.

thomas: This is not how it would look in RDF.

pchampin: I do not share thomas' feeling on this.

pchampin: Please wait until that section is complete enough.

pchampin: PR not ready yet.

thomas: I am waiting, but I want to check how this works with use cases.

AndyS: We can change how often we have meetings if that helps.

pchampin: Let's use the tag "later" for some.

pchampin: Other than those, we can declare we are done.

Action: pchampin to review issues and propose to tag 'later' all but those necessary to finalize report

<TallTed> confirming -- current meeting schedule is every 2 weeks, so next is Oct 1?

<pchampin> https://w3c.github.io/rdf-star/cg-spec/2021-07-01.html#occurrences

ora: would like to discuss the problem raised in our workshop paper

<Fabio_Vitali> can we have a link to the paper?

ora: the uniqueness of triples in RDF-star makes it impossible to model some common PG patterns

Paper: https://www.lassila.org/publications/2021/scg2021-lassila+etal.pdf

bye!

<Fabio_Vitali> @ora thanks

<Fabio_Vitali> bye

<thomas> good bye

<TallTed> ciao

<olaf> bye

<gatemezing> bye

Summary of action items

  1. ora to turn his proposal in #196 into a PR
  2. pchampin to review issues and propose to tag 'later' all but those necessary to finalize report

Summary of resolutions

  1. merge #204
Minutes manually created (not a transcript), formatted by scribe.perl version 142 (Tue Jun 1 16:59:13 2021 UTC).