W3C

– DRAFT –
RDF-star WG meeting

18 December 2025

Attendees

Present
AndyS, doerthe, Dominik_T, enrico, fsasaki, gtw, j22, ktk, niklasl, olaf, ora, pchampin
Regrets
AZ, tl
Chair
ora
Scribe
olaf

Meeting minutes

Approval of last week’s minutes: 1

ora: comments about last week's minutes?

<ora> PROPOSAL: Approve last week's minutes

<ora> +1

<ktk> +0

<fsasaki> +1

<olaf> +1

<niklasl> +1

<AndyS> +1

<lisp> +1

<Dominik_T> +1

<j22> +1

<gtw> +1

<enrico> +0

<pfps> +1

<pchampin> +1

RESOLUTION: Approve last week's minutes

Review of open actions, available at 2

ora: I haven't completed my actions
… but promise that I will work on them over the (upcoming) holidays

pchampin: RDF-INTEROP is published
… Echidna is set up for it

ora: Can we close this action?

pchampin: Let's keep it open until we know that Echidna runs correctly once we merge a PR

niklasl: I have updated the text to use the term for INTEROP

Identifying issues to solve before CR 3

ora: okay, let people comment on the action

ora: What did we hear from the TAG?

pchampin: I got a message from Sarven a minute before this call.
… there are responses on the horizontal review issues

<pchampin> w3ctag/design-reviews#1161 (comment)

<gb> Issue 1161 WG New Spec: RDF 1.2 N-Triples (by pchampin) [Review type: horizontal review] [Resolution: ambivalent]

pchampin: I haven't looked at the issues in detail yet
… but I know that the TAG is concerned with the version announcement
… they don't intent to block anything

<pchampin> w3ctag/design-reviews#1159 (comment)

<gb> Issue 1159 WG New Spec: RDF 1.2 Concepts and Abstract Data Model (by pchampin) [Review type: horizontal review] [Resolution: satisfied with concerns]

pchampin: Sarven's message contains some advice on what the TAG thinks we should do.
… The expectation is that we clarify how the version announcement should be used.
… because it has the potential to break some uses of RDF

ora: Should we add some language to the spec or just let this go?

pchampin: Additional language won't hurt?

ora: Will it help?

<gtw> N-Triples could never be "just concatenated". :\

pchampin: It will help the TAG ;-)

AndyS: Where is the comment about the media type not changing?

pchampin: I had some conversations with Sarven and (?)
… they see that we consider both options

AndyS: You can ask them about CSS versioning.
… Regarding declaring version, we could note in Concepts that, at the moment,
… things are stacked. You should choose the highest version.

ora: Good idea.

ktk: The bnode argument should be used.

ora: Regarding the bnode issue, I have many people made the case that you can just split and concatenate. They don't think about the bnodes!

ora: Volunteers to add some language to Concepts?

AndyS: Yes, me. It should be only one sentence.

ora: What about implementation reports?

<csarven> pchampin: We had two calls on this topic, minutes: 1) https://github.com/w3ctag/meetings/blob/gh-pages/2025/telcons/12-08-minutes.md#design-reviews1159-wg-new-spec-rdf-12-concepts-and-abstract-data-model-github---csarven-1 . 2) will go up sometime this week (placed in the same directory)

<csarven> ^ Some background on other formats.. and why they are not exactly comparable to this situation.

<csarven> (Sorry to drive-by share text.. can't join.. :)

ora: What's the rule? We need two implementations, right?

pchampin: We need two independent implementations for every feature.
… So, in theory, we may have only incomplete implementation.

ora: For each *new* feature?

pchampin: The way it is measured is that the tests pass.

ktk: What does "feature" mean?

TallTed: The purpose is to show that every feature, as defined in the spec, can indeed be implemented.

pchampin: In terms of tests, we can reuse from the RDF-star CG
… We also discussed the EARL reports.

AndyS: Do we know two people who can cover the semantics tests?
… They seem to have been broken since 6 months.
… Apparently, none is running them.
… Who has implementations for semantics tests?

ora: Are you saying these are not serious things?

AndyS: The fix is trivial. It just shows that none is running the tests.

doerthe: I could try to implement something (in January) to be able to run these test.

pchampin: Corese implements the specs and they are currently considering the tests.
… They may not have come to the semantics test, but we can count on them to get to it.

ora: Anything else to discuss regarding getting to CR?

<ktk> ad-freiburg/sparql-conformance

ktk: QLever team build this. It is not specific to QLever.
… It is a test *suite* to run the tests.

AndyS: Good to see that.

<pchampin> I also have one for Sophia, but I need to publish the code, though

AndyS: We (with Thomas and Ruben) are working on the SPARQL tests.

ora: Other business then?

AndyS: Back to TAG response
… (reading comment from the GitHub issue)
… "accidentally"

pchampin: Looks like a misunderstanding.

ora: Any "RDF-star" feature would mean new details in the synatx, which would break old stuff.

AndyS: Text direction as well.

pchampin: Arguably, text direction could be ambiguous.
… Turtle grammar allows for something.

<AndyS> Turtle 1.1: [144s] LANGTAG ::= '@' [a-zA-Z]+ ('-' [a-zA-Z0-9]+)*

pchampin: RDF 1.1 Turtle parser can be expected to still parse it and assume it is a language tag. The grammar allows that.

AndyS: That's not true.
… We are save.

pchampin: Right, I was wrong.

ora: Should we tell TAG this?
… I think we should.
… In addition, should we have some language in the spec, because it may not be obvious to the readers.

AndyS: There is a sentence that kinda covers that.

pchampin: Two things:
… 1) the abstract syntax extends the 1.1 syntax and 2) every concrete syntax extends its 1.1 version
… We can prep some boilerplate that we can use in all syntax-related specs.
… Maybe we should add something in the Changes section.
… I can prepare something.

AndyS: That's an argument for Trig, Turtle, etc, but RDF/XML is different.
… I am not sure what the spec says.
… The processing algorithm might not catch it.

ora: We should check.
… And put some additional language about it in the RDF/XML spec.

<j22> I can open an issue on the rdf/xml to make sure that this is ok

ora: We didn't change the syntax of XML.
… yes, j22, please open an issue

ora: Anyhting else?

Any Other Business (AOB), time permitting

<j22> I opened w3c/rdf-xml#77

<gb> Issue 77 Ensure that the attributes and elements do not cause issues for rdf/xml 1.0/1.1 (by JervenBolleman)

<j22> if you want to trust a good bunchmark have a validated LDBC one

<j22> Want to note that due to high compliance to standards in rdf/sparql world that it is easy to actually benchmark your own workload on many endpoints.

Summary of resolutions

  1. Approve last week's minutes
Minutes manually created (not a transcript), formatted by scribe.perl version 248 (Mon Oct 27 20:04:16 2025 UTC).

Diagnostics

Succeeded: s/hat/that

Succeeded: s/hat/that

Maybe present: TallTed

All speakers: AndyS, doerthe, ktk, niklasl, ora, pchampin, TallTed

Active on IRC: AndyS, csarven, doerthe, Dominik_T, enrico, fsasaki, gtw, j22, ktk, lisp, niklasl, olaf, ora, pchampin, pfps