Meeting minutes
<AZ> Ora: let us put a link to the scribe list in the agenda
<AZ> we did not declare the scribe
Approval of last week's minutes: 1
<AZ> minutes look ok
<ora> proposal: approve last week's minutes
<ora> 0
ora: anyone any worries about last week's minutes?
<gkellogg> +1
<pchampin> +1
<AZ> +1
+1
<niklasl> +1
<AndyS> +1
<Amin_Anjomshoaa> +1
<gtw> +1
<Souri> +1
RESOLUTION: approve last week's minutes
Review of open actions, available at 2
AZ: I added a new message with a proposal for the 'compliance' question
… the proposal contains a "RDF 1.2 Basic" profile
… and "RDF 1.2 Full"
… where the former excludes quoted triples
… Then, for the various specs, there need to be statements saying that "this piece is not needed for Basic"
… Question was how this would affect SPARQL.
pchampin: Regarding concrete syntaxes
… why do we need to say something there, rather than just saying which profile one is in.
AZ: not strictly needed but good for clarity
pchampin: Okay, it would helpful
ora: Are we going to have all possible combinations when SPARQL comes in the picture?
AndyS: I thought we are talking only about RDF Basic and the corresponding effect on the abstract syntax
gkellogg: If I have a Turtle parser, do I have to parameterize it to only emit RDF 1.2 Basic?
… How to communicate that a parser can have two modes?
… What happens if one merges a "Basic" graph and a "Full" one?
… Can SPARQL "Basic" be used to query a "Full" graph?
ora: So, it is indeed more components.
gkellogg: Reasoners are another issue.
… It becomes a combinatorial issue.
… We need to treat lightly.
AndyS: For RDF, we could put a note in all the grammars. That should be practically possible.
… For SPARQL, the challenge is whether we actually need it.
… It gives all different combinations, and it affects the result formats as well.
AZ: For the serialization formats, it may not be necessary. Just for clarification.
… I imagine these two profiles to be similar to the OWL profiles.
… For OWL, there are more profiles.
… and it doesn't make things more complicated.
… Regarding SPARQL, I agree we should keep it simple. Not a "SPARQL Basic"
… but we need some text in the SPARQL spec.
niklasl: I agree with gkellogg that there are a bunch of problems in this.
… A question is, how does "RDF 1.2 Basic" related to RDF 1.1?
gkellogg: text direction may be a difference from RDF 1.1
ora: We don't have any profiles right now. So, we need some baseline. It makes sense to include in such a baseline all those little fixes we are doing.
<AndyS> Current RDF 1.1/1.2 changes: https://
niklasl: Another thing would be that a parser could parse TriG-star into RDF 1.1
… It is necessary to track how "RDF 1.2 Full" is used in the wild.
… You might have a parser that converts TriG-star into reification statements.
<pchampin> even in the CG report, unstar(G) does not have the same meaning as G
AndyS: Shouldn't that better be seen as a transformation.
<gkellogg> w3c/rdf-concepts#34
<gb> Issue 34 Terminology for graph/dataset without quoted triples. (by gkellogg)
gkellogg: We have an issue on terminology here.
… We need terminology to be able to talk about triples that do not contain quoted triples, etc.
… Regarding transformations, there may be no practical way to "round trip"
AndyS: You can do the reverse process.
… How is Issue 34 different?
gkellogg: Such terminology can be useful; in particular, for talking about profiles.
niklasl: Yes that's important.
… I am concerned with backwards compatibility.
… Is the notion of profiles useful for us to distinguish?
… I have talked to people who are worried when I tell them that there will be profiles in RDF.
… I compare the unstar process to how RDF Lists have been handled.
… as syntactic sugar
<pchampin> "lean" also exists, but it is something different
AndyS: There is a notion of well-formed for RDF lists, but I don't think it is officially defined.
niklasl: Yes, so it might be useful to have well-formed reification.
<pchampin> https://
ora: How do the 1.2 profiles relate to 1.1?
… e.g., does supporting 1.2 Basic require supporting 1.1?
<pchampin> gb, status
<gb> pchampin, the delay is 15, issues are on, names are on; and the repositories are w3c/
<gb> … w3c/
<gb> … w3c/
<gb> … w3c/
ora: I realize that this issue is more complication than I initially thought.
Repository: w3c/rdf-star-wg
ora: Should we define the compliance in terms of these difference "components" (parts of the overall set of specs) when they define different profiles?
… I start to think the answer to this question is yes.
pchampin: yes
… Regarding niklasl concerns, the boundary between the profiles should be such that the transition from 1.1 to 1.2 is as smooth as possible (for system that have not (yet) implemented support for quoted triples).
… To give some ground to what kind of compliance can be expected.
niklasl: It is also not fully determined what the RDF-star part exactly is (semantics!)
… generally, it should be made as less complex as possible
ora: other actions?
pchampin: The ones marked as complete can be closed.
… The PRs of rdf-tests appear on the dashboard now.
… No progress on Action 77
… Configutration of Github bot is still a mystery to me.
<pchampin> gb, status
<gb> pchampin, the delay is 15, issues are on, names are on; and the repositories are w3c/
<gb> … w3c/
<gb> … w3c/
<gb> … w3c/
pchampin: I don't know what changes the order of the repos in github bot
… The last action then is about the "last version" of each spec.
… What "last version" means is not fully clear.
… It may refer to the last published version.
… There may be some "accidents" in some of the 1.0 docs.
… I have asked the W3C Web masters
… When they come back, we can decide to fix the ones that are broken.
gkellogg: W3C is inconsistent with the numbering and the popups that point to newer versions.
… we should do both
… it is a question for the documents team, but we should try to be consistent
AndyS: A comment from the community was that the 'latest' link should go to the latest published version, not to the latest working draft.
<TallTed> "Latest version" is ambiguous. "Latest published version" and "Latest editor's draft" would be clearer.
ora: It seems there are two equally natural notions of what "latest" means.
gkellogg: What JSON-LD there is a latest published version, and latest recommendation
<niklasl> +1 for latest published Rec. (highest number with Rec status)
gkellogg: I don't think here is the place to discuss/decide how these things are supposed to work.
… We should call for clarification.
pchampin: Okay, I keep Action 84 open and will get back to the group once the Web masters have replied.
what to discuss atTPAC
ora: AndyS wanted to discuss what to do at TPAC.
AndyS: We have limited amount of time at TPAC and we should use that time well.
ora: We still have two weeks to decide.
gkellogg: There are three major topics that overlap with I18N
… we should have a joint meeting with them
… Having an agenda beforehand would be helpful
… in particular for people who plan to join remotely.
AndyS: In terms of items, the semantics needs to advance. I am hoping that the semantics task force comes out with a proposal.
… Additionally, I would like to schedule some time for the most controversial outstanding issues for SPARQL.
ora: If you have opinions of what should be put on the agenda, mention it on the mailing list.
… We will then put together a tentative agenda.
<pchampin> the label "discuss-f2f" is already available on all our repos
gkellogg: What has been useful in the past is to have topics for which specific people can present.
AndyS: Do we have a time booked with the I18N folks?
ora: I will ping Addison (I18N) (?) again.
<niklasl> ... Is w3c/
AndyS: What part of the day would work best for you?
gkellogg: 12pm until ...
pchampin: Regarding gkellogg's suggestion, should we add a new type of "f2f" tag/label that people can use to tag issues in the repos that they want to discuss during the F2F at TPAC?
<pchampin> w3c/
<AndyS> https://