11 June 2021


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

Meeting minutes

<ora> I never volunteer because Zakim hates me.

<ora> s/not said/said

<ora> I know it is just tough love from zakim...

Announcements and newcomers

pchampin: announcement of newcommers, none

olaf: i did not see mch as eswc. te only one was the same use case inoe of the recent blog posts
… using rdfstr to capture access control statement. on top of graphdb and happy with the performance and withthe conciseness of the expression of the policies

both data and queries

<thomas> which talk was that?

pchampin: also mentioned in the industry panel
… the paper has "rdf-star" in its title.

<olaf> The paper title is "Use Case: Ontologies and RDF* for Knowledge Management"

pchampin: another announcement: completed a rust implementation of the four parsers and intend to publish an earl report

Open actions

<pchampin> https://github.com/w3c/rdf-star/issues?q=is%3Aopen+is%3Aissue+label%3Aaction

pchampin: open actions. 2 on me. 1. no news about the charter. 2. report on media type registrations, that the contact person should be reachable.
… came to the conclusion that there is no issue.

<pchampin> https://github.com/w3c/rdf-star/issues/186

pchampin: reported in the action
… difference btw contact and change controller as the controller remains the w3c
… recommends to close the action

gkellogg: the issue behind this is that with turtle, where the registration has not been updated.
… that w3c is the controller does not provide the actor to actually make the changes

TallTed: we need a path, the w3c role does not suffice

pchampin: push it to the w3c.

TallTed: make that the action

pchampin: we should take the initial action even if it is orthogonal to the group's task

TallTed: suggests to consolidate all requests

<gkellogg> JSON-LD does have W3C as the change controller

TallTed: esp with regard to the change controller

pchampin: can we close that action?

Publishing a new report

pchampin: seeing no other actions, next agenda items.

<pchampin> https://github.com/w3c/rdf-star/pull/185

pchampin: first, thomas' pull request. that some discussion of distinction btw draft and final report.

pchampin: it is large, which is a concern.

thomas: summarizing: it was not intended to be merged. just to give an impression of how to tackle the problem which i see in the draft
… it leaves too much out and does not adequately realted to other things in the field.
… i would like to discuss it in a general way. among reactions was thai i undercut the effort, whcih was not the intention.
… the effort does not do itself a favor if it leaves out the problems which we have discussed.
… the general tone of the report should be more balanced.

<pchampin> q.

thomas: question, is the general sentimet is that it is a "no go" or if it a start to be improved upon.

pchampin: having writte most of the sections which are called "unbalanced", they were my understanding of the balance to this point.

gkellogg: stuck me that "motivations and caveats" is a new section. begins ewith the characterization that the semantics is new and unusual.

but, in itself it does make reference to issues with semantics which are "important to discuss". muchof what is said there would better appear in an appendix.

the majority opinoion has supported "opaque" semantics and controversy is better left inan appendiy, ratehr than in themain body.

<AndyS> s/opionoion/opinion/

thomas: i do not want to questiona design decision, but the problem is stil not clear and needs to be discussed in order to avoid untoward consequences.
… granted, the introduction is too long.
… not questioning the decided dirction, but it needs to presented with more care

TallTed: if i misunderstand "opaque v7s transparent", then your explanation needs more work.
… the current goal is towars a working group to work on rdf star. that need a clear problem statement.
… all work in progress will be problematic, because none are implementing the propsa in the draft. so we need as clear a solution as possible.

thomas: we disagree as to whether the proposed solution actually covers the problem space
… alot of people are implementing under misconceptions

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

pchampin: we do have things like "that" in the current draft.
… there is a remarks section at the ned of the semantics section, whcih contains different things about the consequence sof the semantics.
… esp expalinging the consequences of opacity and describing the alternatives
… the text may give too much credit to the current formulation, but the structure with additional information is there already.
… reluctant to put the qualification at the beginning of the exposition. that does not provide balance

thomas: i could make it shorter and move it into the overview
… the sections to document the discussion remain valuable.

AndyS: the structure of the document is correct.

james: as a reader, I find one point in Thomas' proposal valuable
… the semantics section belongs earlier in the document

pchampin: i do agree that the reorder could be beneficial

gkellogg: speaking for developers. detailed semantics belongs at the end.
… that notwithstanding it is imported to understand what this is, which is not possible without discussing the semantics.
… also important is explainable ai
… but leave alternative semantics ducsissions to secondary locations so as not to confuse the reader

AndyS: there was a strong case make earlier, that people would implement wrongly
… we have tests which should prevent that

TallTed: until the document is finalize, it cannot be implemented.

AndyS: why cannot we turn this into something which is observable

thomas: I think what I do is important, and you will benefit from it too.

thomas: test cases is not my world. that is a techcnical problem not soemthing to discuss

james: I'm sensitive to AndyS's position,
… but ask him to consider Thomas' points

thomas: the concern is that the technical result will break and will benefit no one

pchampin: it _is_ a technical specification and beyond that people can do other stuff
… beyond what is bjectively observeable i am not convinced there is much to do beyond authoring the specification
… i invite other opinions

pchampin: now there are some concrete proposals implicit in the pr.

considering whether to move the semantics up to precede sparql-star looks more consistent

AndyS: in the rdf spec there are separate documents : concept and model theory

TallTed: authored by separate groups

TallTed: in this document there is one group and one document, which argues for purposeful interleaving.

gkellogg: we could procde multiple document, but that is not most useful
… wait for that to get to the working group(s)
… we woud be producing ten different specs, which we could do but the purpose of this groupd to lay out a coherent argument
… it serves better to see that in one place

TallTed: that argues that the order does not matter that much

TallTed: it might be important to label the implementation reports are preliminary in order to avoid misconstrual

gkellogg: do not know what to add to an earl report as it just describes the details of execution

TallTed: general problem with w3c documents as they do not describe their contingencies
… there are large efforts to implement this and customers are looking towards it, but that is irresponsible.

AndyS: does the document's draft status not suffice?

TallTed: it is intended to do the job, but it is not sufficient

AndyS: best taken up outside of this group.

pchampin: there is a w3c community process, ...

ora: the reason why there are multiple rdf specs to a large part has to doe with the work flow and division fo work.

responsibility for documents was contingent on working technology
… unsure how to qualify the document to limit others's activites.

AndyS: we should invite implementation reported from enterprise implementers

gkellogg: concern is that people with look at the final report, where the stauts will say it is neither w3c nor w3c standards track.

<pchampin> +1

gkellogg: we could put something in the processor which generate the html to add qualifications

thomas: wrt gregg's notion of order: semantics is too important to put it at the end.

gkellogg: user community does not demand an account of relation to the semantics.

thomas: if semantics is moved up it would be easier to integrate withthe text.

TallTed: ordering sections: go back to "what is to be solved". more than one. they are not yet clearly statd and eliniated.
… because of that suggestions of how to address them cannot be clearly stated.
… there is more than one problem, but people focus on their own problems.
… people say they want to annotate one triple, but they will actually want to annotate a string of triples.
… that maybe work for anotehr group, but that will come.

<gkellogg> Annotating multiple triples seems to be handled by the N3 group, which has a different theory of operation.

TallTed: re status: qualification, that "htis is not standards" is not adequate, as it turns out that documents are used in more ways that authorized

AndyS: w3c is trying to conform to the trend that efforts form around documents and it is difficult to act against the. this wheter that goes well or not. no working on the odcument will chage the way the communtay clusters and acts.
… other examples are confusing: working group notes end up in tr space.
… we are not exacerbating the situation.

ora: the importance of semantics
… people focus just on concrete things, but that is a w3c failure not to make people appreciate that semantics is concrete
… teds' "groups of statements" : that is a complex issue which is much larger than we can address is this group.

<AndyS> https://w3c.github.io/rdf-star/reports/index.html ==> status section is "This document is merely a W3C-internal document. ...This report describes the state of implementation conformance at the time of publication."

<TallTed> "How to annotate multiple statements" may well be fodder for another group; worth noting sooner than later as "potential work for future WG"

pchampin: today was an important dicussion, but disappointment that we didi not get to discuss what to publish next

<thomas> why not publish a new _draft_?

pchampin: reconvent next week (not bi-weekly) and the first point should be about publishing a draft
… and other quick final word?

ora: we are not done when a clock runs out. were are done when we are done.

pchampin: want to signal progress by publishing drafts

pchampin: thanks to everyone.

Minutes manually created (not a transcript), formatted by scribe.perl version 142 (Tue Jun 1 16:59:13 2021 UTC).