Meeting minutes
Announcements and newcomers
ora: I was at an NSF Open Knowledge Networks meeting, and RDF-star is getting some buzz.
… This is NSFs take on building open knowledge graphs for the public sector/government.
olaf: I talked to people doing machine learning at ESWC to put context on edges from which they want to learn.
… There was a poster paper about access control using RDF-star.
ora: I've had many discussions lately about RDF-star, so people are starting to want to use it.
… What always comes up is that people don't know how to use it for modeling. We need to address this at some point.
… Best Practices, or "RDF-star schema"
Open actions
<pchampin> https://
<pchampin> rdf-star#247
<ghurlbot> Action 247 ping DanBri about publishing the final report on the CG page (on pchampin)
pchampin: I had an exchange with danbri, who appologized and that I had privilages, but only the chair does.
… He suggested I become co-chair. There are precedents for team members to become chairs.
… I think it's on danbri to add me.
… In principle, this would unblock this issue
<pchampin> https://
<pchampin> #43
<ghurlbot> Action 43 organize a call with Adrian Gschwend (on hartig) due 27 May 2022
pchampin: This was about contacting Adrian.
… He is interested, and discussed with colleagues and his company will join W3C so that he can properly chair.
… We'll discuss further in a subsequent adgendum.
<pchampin> #42
<ghurlbot> Action 42 come back to Vladimir (on pchampin) due 27 May 2022
pchampin: I met Valadimir last week in Vienna. Onotext is a W3C member and he is quite open to be an editor or chair.
<pchampin> #29
<ghurlbot> Action 29 Add Ora Lassila as one of the expected chair (on pchampin)
pchampin: I raised the point about not only having triple-store vendors heading the WG.
… Ora has been added, so #29 is done.
<ora> yea!
WG chartering
<ora> (stepping out for one minute)
We passed the penultimate step with W3C Management, and now needs to go to vote by advisory group.
… We need to fix a couple of things: We need to agree on list of chairs, and the timeline.
… There were some questions about that from W3M.
<ora> (back)
gkellogg: there are so many spec, it makes sense to divide out responsibility
… more chairs can be useful, if they work together well
… chairs should usually not be editors, at least not on the items they are chairing
ora: I'm not too excited about having more than two chairs. If Vladimir wanted to be an editor, that would be great.
… Do we have a process for deciding who should be chair?
pchampin: The charter doesn't cover this. I'm a bit nervous about taking such a role in deciding this.
… I asked if it's a problem that it's not complete, which they could live with before submitted.
… I put forward the names on our list.
… If we had no idea, W3M might object.
AndyS: Who the chairs are is up to us. Without that, there's no way to make decisions. Usually W3C comes up with the chairs.
… In terms of numbers, I think there's a slight downside for having too many chairs. If anyone has two roles (editor and chair), I would tend to keep them separate.
… Eventually, it comes down to who has the time for 2+ years.
pchampin: I agree that there is a lot of work, but this calls for many editors, not many chairs. Having too many chairs can reduce the sense of responsibility.
… Let's not try to be too formal about this. I'm hearing that two chairs seems to be enough, and let's try to attract as many editors as possible.
… Ora's already in the charter. (still interested).
… We need to pick between Vladimir and Adrien.
olaf: Given that Vladimir has also expressed interest in editing, that would speak in favor of just using Adrian.
… Also, Adrien is not a triple-store vendor, which is a plus.
pchampin: My understanding was that the role of editor was also interesting.
… I think Ora should have the final word, as he'll have to work with them.
ora: I don't really know either of them, but it seems Adrien might be most appropriate.
AndyS: I'd rather it _not_ be a vote.
Action: pchampin to contact Adrian and add him as chair in the charter draft
<ghurlbot> Created action 46 contact Adrian and add him as chair in the charter draft (on ) due 17 Jun 2022
pchampin: Phillipe was a bit more concerned about the deadline of the first deliverable, which he thinks is too optimistic.
… These specifications are eight years old so will have to go through horizontal review. Given the age of the specs, even with minimal changes, the horizontal review groups may need more time.
AndyS: The SPARQL specs aren't even in ReSpec.
pchampin: I don't think the RDF versions are in ReSpec either.
AndyS: There's a lot of material. While it should be well-laid out, we need to plan on work time.
… We'll need a lot of full and frank discussion about the meaning of RDF-star, which will be time-consuming.
… In the past, I know that people spent time on WGs that were overwhelmed.
ora: 12 months sets an expectation. If we say 2 years, it will take 2 years. Even if people think it's unrealistic, I prefer to set an aggressive expectation.
… I saw that CSS FlexBox has been worked on for twelve years and has never gone to REC.
TallTed: One of the reason of setting target dates is to avoid this very problem.
ora: CSS is hugely successful, but has taken forever.
… Actually Thirteen years
gkellogg: a lot of work happened in the JSON-LD CG before the WG started
… and still there was long discussions in the WG, even though the people involved were mostly the same
… Things take longer than expected.
pchampin: I don't have a strong opinion. Looking back, the documents with a 12-month deadline are Concepts and Abstract Syntax. THen there's N-Quads, N-Triples, Turtle and Trig.
… RDF Schema 1.2 is at 18 months.
… The Turtle family should not take that long. It's basically what's in the 2014 paper and seem to be low-hanging fruit.
olaf: My worry is that people might not accept the Abstract Concepts REC before corresponding Semantics is available.
pchampin: RDF docs range from 12-18 months, and SPARQL from 18-24, because SPARQL comes after RDF.
… All docs for which we have content come earlier, and the others come later.
… Maybe we should have all RDF in 18 months, and all SPARQL does in 24. This could address inter-dependencies.
ora: I think that's fine, but I'd prefer it to be more aggressive.
pchampin: For some people the Semantics come first, for other's Abstract Syntax. Those will likely lead to the most discussion.
<pchampin> proposed: move all RDF specs to 18 month, and all SPARQL specs to 24 months
<gkellogg> +1
<pchampin> +1
<ora> +1
<doerthe> +1
<AndyS> +1
<olaf> +1
ora: All these timelines are speculative anyway.
<TallTed> +1
Resolution: move all RDF specs to 18 month, and all SPARQL specs to 24 months
pchampin: It is speculative, but we don't want the AC to accept or reject due to this.
… The start date will be decided once accepted, and the other stuff in yellow will be resolved later.
… We need a timeline for FPWD.
ora: Given that we have the documents we can set an aggressive timeline for FPWD. Perhaps 6 months.
AndyS: If the SPARQL specs are 24 months, there is no time for review.
pchampin: It depends on what completion means, I took that to mean REC.
pchampin: 6 months for RDF docs, 12 months for SPARQL?
… We don't have the SPARQL timeline totally sequenced behind RDF.
AndyS: I think it could be over-engineered.
pchampin: Phillipe noted that the timeline is pretty arbitrary. It needs to be complete, or removed.
<pchampin> proposed: FPWD after 6 months for RDF docs, 12 months for SPARQL
<pchampin> +1
<gkellogg> +1
<AndyS> +1
pchampin: I plan to send this to the AC next week, and they have one month to vote.
ora: The sooner the AC gets this the better. We've worked on this for some time already.
pchampin: I will make a PR with a tentative timeline for FPWD and we can iterate. I'll send to AC next week.
AndyS: We only need to meet if something comes up.
<olaf> +1
AndyS: I'll send a tentative agenda for the first Friday after the vote.