Meeting minutes
Approval of last week’s minutes: 1
<ktk> PROPOSAL: Approve last week's minutes
ktk: Comunica released a 1.2 implementation. It means we now have 3 implementations of the drafts Jena, Oxigraph and Communica
<ktk> +1
<pfps> +1
<tl> +0 (didn't attend)
<niklasl> +1
<Dominik_T> +1
<lisp> +1
<gtw> +1
<doerthe> +1
<AndyS> +1
<TallTed> +1
<AZ> +0 (wasn't present)
<Tpt> +0 (not present)
<olaf> +1
<Souri> +0 (not present)
RESOLUTION: Approve last week's minutes
Updates from the SPARQL TF
<AndyS> w3c/
<gb> Issue 326 Remaining erratas and errata-like issues (by Tpt)
Tpt: list of errata (link above), none seem significant
Tpt: w3c/
<gb> Pull Request 327 Clarify that unaggregated variables in SELECT raise an error (by Tpt) [spec:bug]
Tpt: another one (link above) is a clarification
… consider "SELECT ?s { ... } GROUP BY ?o" -- some parts of the spec consider it a syntax error
… other parts suggest that ?s should be replaced by SAMPLE(?s)
… there seems to be agreement that it should be a syntax error; most implementations consider it like that
… if we make it a syntax error, then implementations may use it as an extension point
AndyS: I think the spec considers it as a syntax error; only the algebra handles it as if it were not
… we also have an old test for it (rejecting it as a syntax error)
… it is more a matter of getting the wrong impression, depending where you look in the spec
lisp: Is it a consensus that a syntax error allows to be an extension point?
AndyS: You can't change the syntax and call it SPARQL 1.2. Lots of systems have extensions in the syntax but cannot call it SPARQL in the strict sense
Horizontal Reviews Tracking 2 3 4
Tpt: I agree with AndyS. What I meant was : people wanting to extend SPARQL cannot do it if we specify it as a SAMPLE.
pchampin: Now would be a good time to ping the horizontal review groups
Roadmap to completion in 2026
ktk: What do we want to send next to horizontal review?
pchampin: We might split the Turtle family of syntaxes from there (RDF/XML and schema)
pchampin: We need to reflect the changes in Turtle to the other formatss (the profile would make sense to RDF/XML)
pchampin: We are slightly closer to completion on Turtle/TriG/NQuads than the other two
pchampin: It might make sense to try make those 3 out there first
AndyS: What we can start now is doing group reviews of the documents
AndyS: To see if there are issues in the spec
<pchampin> +1 AndyS
ktk: Is there something with a big list of TODO?
Tpt: When RDF terms and EXISTS support is done I feel we can push SPARQL 1.2
Tpt: we can fix the remaining issues later
pchampin: I feel there are still big issues on property paths for example
pchampin: Notes have a much lighter life cycle
pchampin: But it's not that they are just not that we should lose track of them
olaf: There two aspects to it: the spec on property paths still have a bunch of smaller issues
olaf: indeed there are no formalization of triple terms on subject/object position of triple path. We have a PR in progress
ktk: Is there any big issue remaining in property path?
Tpt: I am not aware of any issue affecting implementation, it's only formalization issues at my knowledge
olaf: We are not planning to add any new feature on property paths at the moment like p{n,m}
ktk: We are pushing for NQuads/Turtle/TriG as the next batch
ktk: We are not pushing new features to SPARQL but focusing on stabilization
ktk: Does anyone agree?
AndyS: What about RDF/XML?
ktk: We should catch up with RDF/XML editors
<niklasl> w3c/
<gb> Issue 80 rdf:annotation and rdf:annotationNodeID may be confused by pre rdf/xml 1.2 parsers (by JervenBolleman)
pchampin: I agree RDF/XML is properly in good shape. I realized only recently that new features in RDF/XML 1.2 won't be rejected by 1.1 parsers
pchampin: I did not realise that RDF/XML did not share this nice feature
AndyS: One of the features that is its:dir that we inherit from the its spec is enough to change the parsing
AndyS: The rdf: namespace is effectively reserved. It is on the group control. We can hide behind "you should not mess with it"
AndyS: We can't easily change the syntax to make sure the parser will break, opposite to Turtle
pchampin: I agree regarding its:dir. Reserving the rdf: namespace post hoc is strange
AndyS: There is already some note on it
AndyS: There is an issue/PR on RDF/XML discussing it
pchampin: I just wanted to make sure we make an informed decision about it
ktk: What is the best case scenario for the publication process? How many months?
pchampin: The more variable step is the horizontal review. We might be very clear that we give the same answers for every RDF spec
pchampin: The AC review is at least 28 days
Review of open actions, available at 5
<pchampin> nothing to report
Identifying issues to solve before CR 6
pchampin: w3c/
<gb> Issue 118 explain how to interpret multiple `VERSION` directives (by pchampin) [ms:CR]
pchampin: w3c/
<gb> Pull Request 90 adds a paragraph stating that N-Triples 1.2 is not ambiguous for 1.1 implementations (by pchampin) [ms:CR]
AndyS: I have an MR to centralize the version wording in concepts: w3c/
<gb> Pull Request 261 Upwards compatibility of versions (by afs)
<ktk> w3ctag/
<gb> Issue 1161 WG New Spec: RDF 1.2 N-Triples (by pchampin) [Review type: horizontal review] [Resolution: ambivalent]
ktk: ^ comments on horizontal reviews
AndyS: I made a PR to remove closed issues from a document. Do we keep or remove the issues appendix if it's empty?
AndyS: I think we should remove it because it's for the WG internal work
AndyS: I will update the PR
<pchampin> I think it would make sense to leave it in WDs, and remove it in the CR snapshot
Any Other Business (AOB), time permitting
<ktk> https://
ktk: I will close these ^ two issues later
AndyS: We have a SPARQL meeting tomorrow
ktk: Have a good day/evening!
<ktk> s/unagreggated variables/unaggregated variables/
<ktk> s/there seem to be agreement/there seems to be agreement/
<ktk> s/can not do it/cannot do it/
<ktk> s/Now would be good time/Now would be a good time/
<ktk> s/from the there/from there/
<ktk> s/slightly to completion/slightly closer to completion/
<ktk> s/To see if there issues/To see if there are issues/
<ktk> s/live cycle/life cycle/
<ktk> s/loose track/lose track/