Meeting minutes
<ktk> start meeting
Scribe: Haudebourg, Timothée (alternate: Arndt, Dörthe)
<TallTed> gkellogg has volunteered to scribe
Approval of last week's minutes: 1
AndyS: Perhaps we can get updates from task forces?
ktk: We don't have enough people present to approve minutes this week.
Update on Semantic TF (unless Enrico sends regerets)
pfps: We're having informal meetings with minimal logging in the semantics TF.
… There are a number of semantic proposals, and Enrico has an outline of how to characterize them.
… They're in the rdf-star-wg repository, but unofficial.
… The semantics TF is waiting for decisions on which use cases will be supported, so we know how they should behave.
… The default is that the semantics would be have in away consistent with the use cases.
… The TF semantics don't seem to work well.
<AndyS> WIP semantics on the wiki: https://
pfps: Building true semantics can be a pain, as there are several choices to be made, which aren't clear until examined closely.
… It's worth looking at the proposals.
… it's easier to write down in a cryptic/terse fashion, and the consequences become documentation.
AndyS: THere's been an ongoing difference between syntactic and semantic notations.
pfps: Good question: can you have a catholic semantics that covers the use cases?
… It's more difficult to have a semantics that covers multiple possibilities. Do we want a single simple semantics or one which has more possibilities?
… This would have implications for the syntax, which could be problematic.
AndyS: I thought we might end up with two syntaxes, but the worry is how to explain it to people?
pfps: We already have different semantic regimes.
AndyS: We can discuss different syntactic variations, but it may be difficult for a broader audience to understand.
TallTed: By "catholic semantics", do you mean "universal"?
pfps: Yes, but I use "catholic" to cover all _reasonable_ alternatives, not all _possible_ alternatives.
… The other thing with discussion is the difference between quoted triples and singleton graphs.
… There's been a bit of discussion on what should be done. The WG needs to say something, if not just, we're not doing this.
… I've looked at the new N3 semantics, which I don't fully understand.
… If quoted triple semantics can be made compatible with N3 is unknown.
… You can think of a quoted triple as an N3 included graph with a single triple. It might be nice for them to behave the same way.
AndyS: It's a tension between talking about claims and the representation. Is it meaning, or pure quotation?
pfps: The thrust of N3 has been that included graphs are used for implication, but nothing prevents them from being used differently.
ktk: I believe doerte said that the goal is to make N3 compatible with RDF, in some way.
pfps: The question is if there is any reasonable way to make it possible.
Review of open actions, available at 2
ktk: no one with an assigned action is present.
… Ora has something to present next time.
Review of pull requests, available at 3
pfps: There's been some discussion on whether the base direction contains more than one thing on it, and perhaps the change in direction needs to be pulled out.
<pfps> currently whether there is one node or two resulting from "hi"@en and "hi"@EN is underspecified
pfps: the issue is if there is a change of the treatment of language tags in the PR.
… It's unspecified if "hi"@en is the same as "hi"@EN.
AndyS: The suggestion is to take that change out of the PR, as it's not related to base direction.
… It's around line 740 in the PR.
<Souri_> Our current implementation considers "hi"@EN as equivalent to the normalized form "hi"@en.
AndyS: The old text says equal character-by-character, and the new says to normalize.
… At the moment, the specs say the en and EN give different terms. This goes back to RDF 1.0.
<pfps> the relevant text in RDF 1.1 Concepts is "A literal is a language-tagged string if the third element is present. Lexical representations of language tags MAY be converted to lower case."
AndyS: The reason they're the same value, is because RDF entailment uses this.
TallTed: The issue seems to be that there's a change in implementation behavior if conforming to the new vs old language.
AndyS: The text in the PR around language tag equality is still somewhat complicated.
… Language tags are otherwise described as "equal values".
TallTed: In the example, the values compared are the same integer, but not the same term.
ACTION: gkellogg to separate out language tag from base direction in w3c/rdf-concepts#48.
gkellogg: the editorial ones can be merged at this point
… N-Triples 34 can be merged probably next week
w3c/rdf-n-triples#34 can wait on w3c/rdf-concepts#36 to be merged.
<gb> CLOSED Pull Request 36 FPWD publication snapshot. (by gkellogg) [ms:FPWD]
<gb> Pull Request 34 Adds BNF and text to extend LANGTAG to support base direction (by gkellogg) [spec:substantive]
AndyS: I didn't get to the media-type changes quite yet. Waiting on Ruben.
AndyS: The ones on the space character needed some resolution.
gkellogg: may need privacy considerations on semantics and schema.
Any Other Business (AOB), time permitting
gkellogg: we are almost done with RDF C14n normative bits, there are remaining editorial changes.
<TallTed> for them what care, https://
gkellogg: might be ready after TPAC
gkellogg: normative parts of RDF Dataset Canonicalization are done. Should go to CR after TPAC.
<ktk> tnx
<ktk> few issues, will see with pierre-antoine to fix them. open actions twice, scribe change not in topic
<ktk> other than that it looks good to me
<ktk> TallTed: thanks!
<TallTed> <TallTed> I have trained gb (nee ghurlbot) to use our repositories. These will stick until/unless the bot is renamed again, or its admins drop its settings file(s).
<gb> I created issue #30
<gb> but I could not add the "action" label.
<gb> That probably means I don't have push permission on w3c/sparql-results-json.