RDF-star WG meeting

Approval of last week’s minutes: 1 2 agendabot]

<pfps> minutes look good, as far as I can tell

ktk: two minutes; one from last week, and one from TPAC

pchampin: RSSAgent was disconnected for a few minutes during the TPAC meeting. I copied local irc log and added what was missing.
… already fixed in minutes that are published.

<TallTed> +1 approve minutes

AndyS: after this can we get insight from pchampin and ktk about TPAC?

<AndyS> +1

<pchampin> +1

<pfps> +1

<ktk> PROPOSAL: Approve last week's minutes (TPAC & regular)

<niklasl> +1

<gtw> +1

<pfps> there is some background noise on the call

<ktk> +1

<TallTed> next Thursday is US Thanksgiving ... question of whether meeting will have enough international attendance

<olaf> +1

<pfps> I expect not to be available next week

<tl> +0 wasn't there

RESOLUTION: Approve last week's minutes (TPAC & regular)

<pfps> ... so I can't be alternative scribe

pchampin: TPAC overall was really good.. Interesting discussions. Very few from this WG.
… WG meeting was mostly for benefit of observers. Lots of questions. ora gave presentation on where we are.
… discussed status of SPARQL.
… we had briefly two discussion topics focused on next steps. One was idea about changing the basic [?] algorithm.
… Enrico had point about improving narrative around propositions in Primer.

ktk: nice to have people from outside the WG.
… highlight was verified credential group. briefly talked to Manu Sporny. Presented about California drivers license using verified credentials in QR code.

TallTed: interesting part from my perspective is QR code encapsulates an encrypted/signed data that is on front of license.
… this is associated with TrueAge project with is verifiable credentials-based. Need to present some hardcode proof abotu identity on first visit. Then streamlined afterwards.
… rolled out in millions of stores needing to check IDs.
… expected to recognize lots of forms of IDs. This is benefiting them in a big way. Bringing RDF into usage in places that nobody expected in these numbers.
… 100M instances of these stores. 50M drivers license in CA. Things in surprising places.

ktk: TruAge is by whom?

TallTed: the convenience stores themselves.

<pchampin> Connexxus

ktk: you don't disclose your birthdate, just prove you are above the threshold.
… "selective disclosure".
… lots of use-cases.
… you can scan QR code on phone. They can revoke licenses as well.

pchampin: verifiable credentials are JSON-LD signed documents. QR codes are CBOR-LD. Not yet a spec, but on charter for JSON-LD WG.
… RDF quads in compact form in limited space for QR code.

<pchampin> https://www.w3.org/TR/rdf-canon/

pchampin: makes use of RDF canonicalization.
… the way digital signatures are added to JSON-LD document is in theory generalizable to other documents.
… not just VC format.
… Ivan H has done some experiments on other formats.

<TallTed> https://www.dmv.ca.gov/portal/ca-dmv-wallet/truage/

<TallTed> https://www.mytruage.org/about/faq

<TallTed> https://www.google.com/search?q=truage+qrcode

pchampin: Web of Things WG. Using RDF and JSON-LD to provide generic descriptions of IoT devices. Lot of interest in what the group presented.
… Retail store association behind TrueAge behind Web of Things.
… demo on automated till interacting with connected objects.

ktk: next week. Thanksgiving?

I won't be here next week.
… don't remember what we did last year.

<pchampin> I'm actually at risk for next week, though that's not thanksgiving related

TallTed: we did skip last year.
… ora will also probably be out.

ktk: I'm in a tough timezone.
… propose we skip next week.
… OK. Skip next week.

<ktk> 1?

Review of open actions, available at 3

ktk: any updates from pchampin or niklasl on issues?

niklasl: not any update on my note. waiting for feedback.

ktk: from the group? or someone in particular?

niklasl: the group. Ted commented it should be moved to another repository.

pchampin: #179 can be closed.

<gb> Action 179 find a way to extract a dependency graph of our specs (on pchampin) due 2025-09-25

pchampin: haven't updated the graph yet.
… #55 still pending. will discuss later on, but need group decision to publish interop.

<gb> Action 55 set up echidna in all RDF&SPARQL WG GH repositories (on pchampin) due 27 Apr 2023

Updates from the SPARQL TF

AndyS: look at the rest of the GH board?

ktk: that is after this topic.

AndyS: taskforce in last meeting had discussion about whether it would be useful to take WG telecon time as things wind down in other documents.
… as we are waiting for review to come in.
… it would be useful to bring things to wider group attention.
… not substitue for the TF meetings which are more detailed discussions.
… good to bring things back to this group for awareness.
… There are a few issues that might be of interest. Olaf leading work on EXISTS.
… There is a discussion of VERSIONs in SPARQL. Whether to put versioning in SPARQL namespace document.
… Consequences on details how it impacts service description.
… Finishing up stuff on triple terms. Making sure all material is in document. Whether things should be sortable/comparable.

ktk: Ones interested could have a look at sparql repo?

olaf: Discussion came up yesterday might be something. Maybe we can first discuss in the group. Things related to property paths.

AndyS: I think that was more of a question.
… Final comment I'd make is we are going to have to prioritize. Not everything possible in the WG timescale. Identify issues that are necessary to address.
… maybe push things out into the maintenance/new features phase.
… We need to show progress before we could justify any extension for SPARQL.

ktk: link to repo?

<AndyS> https://github.com/w3c/sparql-query/pulls

ktk: that was one question at TPAC. What is status of SPARQL?

AndyS: W3C process tends to encourage getting review.
… we need to get through taking in incoming. Some of the things we're getting are just questions.
… we've got to balance with making PRs.
… then the reviewing. Particularly design review. Looking at PRs. Across documents, not just specific areas of the change.

Identifying issues to solve before CR 4

No security & privacy considerations in RDF-Schema

ktk: I was working on making the request for reviews from i18n, privacy, security, accessibility.
… I realized we might have an issue. In RDF Schema, there is no privacy/security section.
… reviews for that say don't submit without those. Is this something we have to add?
… In Schema we did it.

pchampin: the document that is missing them is Schema, right?
… not part of first batch sending for review.

ktk: we had Concepts, Semantics, ...

pchampin: and N-Triples.

ktk: I will have to adjust that.

pchampin: That is going to be an issue soon.
… We probably want to solve that.
… low hanging fruit would be to refer to corresponding sections in Concepts.
… something similar in JSON-LD. Most considerations in JSON-LD syntax document. API points to that.

<ktk> https://w3c.github.io/rdf-semantics/spec/#privacy

ktk: that's what Semantics did.

pchampin: it's a fair thing to do. Don't think there are specific security or privacy considerations to Schema that do not apply to RDF in general.
… need to address, but we have an easy way out.

ktk: create an issue in Schema?
… I will do that.

w3c/rdf-n-triples#73 and w3c/rdf-turtle#89

ktk: proposal to close "IRI resolution".

AndyS: I checked N-Triples, -Quads, TriG. Text has been updated to something reasonable.
… action could be closed for line items 1 and 2.

<pchampin> +1 to close them

ktk: propose we close them.

AndyS: will do it after the meeting.

w3c/rdf-concepts#248

AndyS: on the editors notes one, there is a related PR. #256. Talked about last week.

<gb> Issue 256 not found

AndyS: removes the editors note about conformance. Discussion had stopped.
… I assume that means it's decided. Will merge unless I hear otherwise.
… line item 4.
… does not close line item 4. Three others to do.
… one of which is interop.
… Editing needs to be done when Interop is available to be linked.

<ktk> w3c/rdf-concepts#256

<gb> Pull Request 256 Remove editor's note about conformance levels (by afs)

pchampin: regarding Interop, I believe we discussed that earlier. Could not find a resolution.
… required to initiate first publication.
… to summarize, I think we can publish it as FPWD. Incomplete, but that's why it's a draft.
… non-normatively referring to it, so that's sufficient.
… I propose we make a resolution to publish RDF Interop as FPWD.

ktk: Can do that later today.

<pfps> sound issues

<pfps> i'm about line item 5

<ktk> PROPOSAL: Publish RDF Interop as FPWD

<pchampin> +1

<TallTed> +1

<gtw> +1

<olaf> +1

<ktk> +1

<niklasl> +1 (I have a term choice issue but that shouldn't block)

<enrico> +1

<doerthe> +1

<AndyS> +1

<lisp> +1

<tl> +0

RESOLUTION: Publish RDF Interop as FPWD

<pfps> line item 5 depends on whether Semantics is normative wrt what Schema vocabulary is of interest or Schema is, historically it has been Schema but that makes a normative dependency from Semantics to Schema

w3c/rdf-semantics#163

pfps: Line item 5 is about whether semantics should depend normatively on Schema.
… historically it's been Schema that has been normative.
… issue is whether that's a good idea. Should it be changed?
… that is a WG decision.

ktk: Opinions on that?

AndyS: it's marked "proposed closing".

pfps: that was before pchampin ... [?]

pfps: I marked for closing. pchampin said something afterwards.
… I'll take that tag off.

pchampin: The RDF Schema defines the vocab. RDF Semantics provides some formal interpretations of the terms. Those terms have consequences as inferences.
… two sides of same coin.

<pfps> I'm voting conservatie on this - keep Schema normative - a change to informative should result in lots of informative marks in Schema

pchampin: If somebody has a problem with inconsistency between prose and formal description. I would be more comfortable with formal description having precedence.
… other reason is we're trying to publish semantics before schema.
… having dependency other way around would slow process. Not a good reason by itself, but other reason is a good one. This strengthens.

ktk: shall we vote about it? pfps, WG decision would take the form of a vote?

pfps: yes.

pfps: I'll write something which would make the change.

<pfps> PROPOSED: Make the RDF Schema document informative wrt what vocabulary is of interest and what its meaning is.

AndyS: want to talk about Turtle, not Schema. After proposal.

pchampin: I have an issue with pfps' proposal. Seems to apply to Schema as a document. Not what I was suggesting.
… not suggestion to make Schema a non-normative document.
… suggesting to not normatively depend on it.

pfps: problem with that is we have 2 normative documents as to Schema. There are Semantics and Schema. Both define things.
… which is more authoratative?
… Could keep Schema normative and say if conflict, defer to Semantics.

pchampin: that's what I was suggesting.

AndyS: exact wording of proposal I read as making whole of Schema a Note.

pfps: would rip out all normative stuff in Schema.

AndyS: framing of proposal would make it a non-normative document.

pfps: what would be a smaller step to remove normative link from semantics to schema?

AndyS: I thought idea of saying semantics document is authoratative would be sufficient.

pfps: somebody else needs to craft that proposal.

pchampin: if anything, Schema should normatively refer to Semantics.

pfps: as long as there's conflict resolution statement added, that would work for me.

<tl> looks good

<pchampin> PROPOSAL: change RDF-Semantics to *non-normatively* refer to RDF-Schema; RDF-Schema should then normatively refer to RDF-Semantics (giving precedence to RDF-Semantics in case of conflict)

<pchampin> +1

<TallTed> +1

<niklasl> +1

<tl> +1

<olaf> +1

<doerthe> +1

<pfps> +0.5

<ktk> +1

<enrico> +1

<gtw> +1

<AndyS> +1

<lisp> +1

RESOLUTION: change RDF-Semantics to *non-normatively* refer to RDF-Schema; RDF-Schema should then normatively refer to RDF-Semantics (giving precedence to RDF-Semantics in case of conflict)

pfps: when do the notes/issues saying it was discussed show up? afterwards?

pchampin: yes.

pfps: not in issue yet, but expect to see it soon.

pchampin: I may need to patch minutes. Will do what is needed.

<Zakim> pfps, you wanted to say that there needs to be a conflict-resolution statement

AndyS: relationship between semantics and turtle. I don't mind which way it's decided.
… if turtle suddenly changed, it wouldn't really change semantics.

ktk: what do you propose? we should vote for that?

AndyS: whatever the editor wants.

pfps: I'm uncomfortable making link non-normative. In places where the semantics is defined. Includes turtle constructs.
… if you don't understand how turtle works, you can't understand definitions.
… little bits of turtle. not a lot of it, but curies are included, so can't just include n-triples.
… prefixednames.

pfps: I am uncomfortable making the link non-normative.
… it is pushing the limit. Don't want to have somebody come back to say we have to change this.

<Zakim> AndyS, you wanted to talk about the turtle dependency after the proposal

AndyS: propose cancelling next week's SPARQL TF meeting.