28 May 2021


AndyS, gkellogg, james, ora, pchampin, TallTed, thomas
olaf, william
gkellogg, pchampin

Meeting minutes

<AndyS> Don't forget - zoom call today.

Reminder that we’re on Zoom this week.

Announcements and newcomers

pchampin: Olaf had tweeted about an interview on using RDF-star, but haven’t seen a reference.
… Using Ontotext GraphDB?

<AndyS> https://www.synaptica.com/rdf-star-graphdb-graphite/

<AndyS> Also - there was a blog from TopQuadrant.


pchampin: I think we’re narrowing the things we have to discuss, and it’s been more convenient to have calls less frequently. I think we can go to one call every two weeks.

<james> +0

AndyS: The next objective is to get a community report done.
… I’d think the intro and the SPARQL bits are what remains, both nearlly done.
… When do we plan that for?

pchampin: Maybe we can defer the change of schedule until after the report is published.

thomas: GitHub has been active lately on semantics and vocabulary. Is there interest in becoming more concrete on semantic extensions?

pchampin: My understanding was that the vocabulary was not blocking for the publishing of the final report (first).
… THe idea is to make it a bit more milestone-ish.

thomas: I feer it then becomes an afterthought and is left to the WG. I think the extension part should be as concrete as possible.

pchampin: I made a proposal following my action about what issues are considered to be blocking before “Final Report” milestone.

thomas: I’m fine with “milestone”.

pchampin: This was a list of blocking issues, and the vocabulary was not part of it.
… My understanding is that we can discuss the vocabulary afterwards and have other milestones before a final final report.
… It’s good to show progress before an upcoming charter draft.
… I’m not proposing to stop our work after the FR is out. We have some open issues, and my goal is to address them.

thomas: It’s the fring bits of RDF-star that are left, and intrest will deminish.

ora: I think the questions of semantics are very important. ONe of the big mistakes with the original RDF spec is that we weren’t very formal, which was rectified later.
… I’m in favor of working on semantic questions, but my feer is that we might want to address issues of schema.
… At this point, it’s not clear how to write a schema that makes use of RDF-star.

pchampin: Regarding the vocabulary, it started with an idea to have a term for the “class” of embedded triples, so that we could describe a property to expect and embedded triple.
… We also had the idea of having a standard way of expressing that something is an instance of a triple in a graph.
… There was a question of what voculary to use, but the key was an IRI expressing the type of an embedded triple.
… This opens a can of worms; we agreed not to decide where to put the vocabulary, so this is not low-hanging fruit, which is why I preferred to defer until after the next milestone.

thomas: THe two properties “occurance of” and “in graph” could be defined informally. THat would show how it can/should be done.

<pchampin> https://w3c.github.io/rdf-star/cg-spec/editors_draft.html#occurrences-example

pchampin: The “in” hides it’s own issues. Is the object a graph, or a graph document, …

thomas: I think it can be solved pragmatically.

pchampin: I’m not sure something ambiguous is better than nothing.

ora: I wanted to say I agree with PA, but we have to go there soon. If not now, at least we should have some words to say that something like this is needed.
… I don’t want to end up like the PG community, where there is no schema language.

gkellogg: we have to be careful about talking about graphs.
… RDF does not really provides way to talk about graphs,
… because the semantics of named graphs is not defined.
… Talking about the default graph is sometimes done with <>.

thomas: If it can be done in SPARQL, that’s good enough for me.

pchampin: Can we make such a comitment for RDF?

thomas: We could use mild terms to suggest that it could/should be done.

gkellogg: some groups (VC) depend on the ability to use the graph name to identify it
… If we need to define that, then it needs to be in the charter.
… Certainly will attract people.

pchampin: Back to the final report, I propose to add a chapter on the RDF-star vocabulary, but only reference the open issues plus some non-normative notes about being addressed in the future.

STRAWPOLL have a "placeholder" section about the RDF-star vocabulary, containing links to the issues

<gkellogg> +1

<thomas> +1

<ora> +1

<AndyS> +1: Many issues raised are wider than RDF-star. A placeholder seems the way to go. (Separate RDF-DEV group?)

james: I’m perplexed as to the reluctance for using SPARQL expressions when so much work has been done. Even if you don’t know the consequences, it would help the discourse.
… It would be worth to have this discussion in the chapter.

pchampin: I’m reluctant to make it a blocking condition on the milestone final report. After it’s published, I think we should address it.

james: Then why not put it in now? I’d like to make it as concrete as possible ASAP.

pchampin: My feeling is that it may draw us into discussions that will further delay the milestone.

<pchampin> +1

<james> +1

<TallTed> +0

<thomas> how many "final reports" are there? what differentiates a first "final report" from others? will they be marked as molestones?

pchampin: I don’t know how many, just that we don’t exclude the possibility of having another one in the future.

thomas: I see the time pressure, but if it’s named “Final Report”, I don’t see how we have several of them.

gkellogg: We’re bound by the set of possible W3C document types, “draft” and “final”.

thomas: We could try to get the vocabulary right as James or I have proposed for the next 2-3 weeks. If we don’t get agreement, then we can defer, otherwise try to include.

<james> nb. i am not even proposing that we get it "right", just the the document indicates some application of the terms

thomas: Can we agree to work on it over the next three weeks?

pchampin: This work is a side-project for all of us, so we don’t have strict deadlines anyway.

AndyS: We can have multiple final reports, or split it into parts. Delaying on this point ...
… I think it will interest other people, and we had an agreement on the process, but if we keep changing it, I lose confidence.

pchampin: Our first priority is to get a report out, and wanted to mention the section even if it’s not complete.
… It’s a question of what the editor’s prorities are in the coming weeks. Merging the current PRs or working on a new PR.

thomas: I feel that more people want to postpone it, but I feer that it will be postponed for a long time.

pchampin: That’s reallly not the goal, there is a section that will need to be filled.

STRAWPOLL change to a call every 2 weeks start now

<pchampin> +1

<james> +0

<ora> +0

<gkellogg> 0

<TallTed> +0

<thomas> -1 for a few more weeks

<AndyS> +0.8

<thomas> i could change my vote to -0.5

<pchampin> let's postpone that decsion

Open actions



pchampin: PR #121 has a lot of work done.

gkellogg: some wordsmithing issues left, but I'm astisfied with the substance
… I have always found SPARQL equality challenging
… I'm passing all the tests.

AndyS: There are some more minor changes needed.
… It’s important to have a stable ordering.


pchampin: Next was #1164 on Sparql Service Description. I’ll reference that in the TBD section.


pchampin: The section on the rationale for transparence/opacity. I noticed that PRs fail on CI, while AndyS’s succeeds.

AndyS: I avoided a lot of ReSpec features and just wrote raw HTML. I didn’t have the energy on how to drop in.
… In addition to ARIA, there’s an empty-ID warning.

pchampin: To be clear, the idea is that the different semantics proposed in the document could be seen as layers of semantic extensions on top of the first one.
… I didn’t think that would cause too much discussion,

thomas: I saw this as marketing. I made a mistake in an answer. I had some things mixed up. I still don’t like what we’re doing.
… You’re leaving out common-sense entailments.
… I’m not as concerned about it now, but I’d be much happier if we could point to something about semantic extensions, but we don’t have that now.

pchampin: I don’t want to have the debate about which is best, because we agreed to say that there was no concensus in the group. You might use this as a semantic extension, or as a replacement, but we couldn’t agree.
… The term has a precise definition in the RDF spec. The alternative sematncis work on top of the opaque semantics.

thomas: I’m refering to my discussions with olaf.

pchampin: I’m just try to get to the point where we can merge the PR.
… We’ll need to continue on GitHub. The goal was to see if we can make progress.



pchampin: The charter hasn’t progressed.
… We haven’t done anything on the media-type or conneg discussions yet.
… AndyS mentioned that it could break some things.

james: WRT #175, AndyS requested that it introduce an issue in the sparql-12 group, which I have.

<AndyS> james - thx

thomas: Any feedback on my proposal?

pchampin: I see where you’re going. Having an alternative syntax for transparency is interesting, but until we’ve agreed on the vocabulary, I think it’s too early.
… I think the issue raised by pfps on non-obvious ramifications is a bit embarassing,
… It is dependent on our agreeing on how to add transparency.

Minutes manually created (not a transcript), formatted by scribe.perl version 127 (Wed Dec 30 17:39:58 2020 UTC).