W3C

RDF-star WG

19 January 2023

Attendees

Present
AndyS, AZ, Doerthe, Dominik_T, enrico, gtw, ktk, olaf, pchampin, pfps, Souri, TallTed
Regrets
ora, gkellogg
Chair
ktk
Scribe
AndyS

Meeting minutes

TIL -- There's a list of all previous meetings : https://www.w3.org/services/meeting-minutes?channel=rdf-star&num=200

someone ought invent a standard for for news feed on the web.

<ktk> additional items proposed for agenda:

<ktk> * GH Project for collecting issues [1]

<ktk> * Suri: Slide on storage, performance, and future-proofing aspects related to multi-edge handling (5m)

<ktk> * Gregg: Editor assignments

<ktk> https://www.w3.org/2023/01/12-rdf-star-minutes.html

ktk: accept the minutes
… feedback on agenda
… GH repos
… a GH repo for issues of all docs
… Souri would like a 5 min slot
… gkellogg asked about editors

pchampin: created a GH repo per specification
… github pages active
… 'main' protected

<ktk> AndyS: What is published to GH pages, is it the whole repo or a particular subdirectory?

<ktk> pchampin: it's the whole repo

pfps: Q about how GH page work
… respec can run locally?

<ktk> AndyS: yes it can be run locally, it's some javascript in the spec itself

<ktk> pfps: can we have pre-commits for checking things on commits?

<ktk> AndyS: there are some pre-commit hooks but they only check basic things, not respec related

pfps: will dig into checking possibilites for reSpec.

<ktk> AndyS: I haven't pushed any content on SPARQL stuff yet, I was working separately.

<ktk> ... I got some respec errors and will push after that.

Issues project

ktk: Idea is a GH project for organising work.
… also issues across repos

<pchampin> w3c already uses GH projects, so it should be possible

ktk: one central place to see all work

pfps: New feature of sub-projects may be useful
… "task lists"
… in an issue link to sub issues

<Zakim> TallTed, you wanted to ask how GitHub projects handle GitHub notifications

TallTed: GH creates features that are not fully thought out.
… how does GH handle notifications?

ktk: subscribe is per project
… should be cautious in using the features

ktk: PA and chairs will come back with a proposal.

Editor assignments

ktk: gregg gave a list of docs he's interested in.

<TallTed> +1 handle Editor assignments in an issue. Possibly with checklist of docs at top, to be supplemented with Editors as volunteered/assigned.

<ktk> AndyS: I assume we do not do a complete review-cycle since the 1.1 specs

<ktk> TallTed: I think most changes would be between 1.1 and today. but it would be good if reviewers still have a look at the whole document.

<ktk> TallTed: Errata in a previous edition should be corrected in a new addition

<ktk> ... if we find an error in a previous edition it should be corrected in any way.

<ktk> AndyS: it depends on how much work this creates. we have only 18 months of time

<ktk> TallTed: if errors create work, we should be able to extend the timeframe

<ktk> pfps: if we get a comment that creates massive changes for everything, we can say it is out of scope.

pfps: we have the option of declining comments

<ktk> TallTed: if the need for change is valid, like they discover a broken thing in the current spec, it should be treated with respect.

ktk: Will create issue to record editor offers

ACTION: ktk, to create an issue for editors offers

ACTION: ktk to create an issue for editors offers

<ghurlbot> Created action 11

souri: present slide

The scribe asks that the slide be sent to the WG email list.

<pfps> I'm confused. What is supposed to be happening here? An example that is part of a worked-out use case seems indicated.

timestamp

<pfps> timestamp is the use case? How is this being represented?

I'm confused.

pfps: This depends on what is presented and how. What's the UC?
… the WG should pay attention to use cases
… provenance, modal logic. Can't future proof everything.

<pfps> This looks like an attempt to change the RDF semantics from set-based to bag-based. This is a fundamental change.

TallTed: slide implies an automatic id - which is not always the case.

<AZ> There's existing work (quite a lot actually) about mapping relational data to RDF (and back). It works. Do we need a new RDF that mimmics relational data so much that it's essentially doing the same thing?

TallTed: I think you are asking for implicit and explicit naming
… need to explain the slides

<TallTed> +1 to explicit use case (which I don't have ... yet)

<AZ> +1 for concrete use cases

<pfps> I agree that introducing things via email is better.

Summary of action items

  1. ktk, to create an issue for editors offers
  2. ktk to create an issue for editors offers
Minutes manually created (not a transcript), formatted by scribe.perl version 197 (Tue Nov 8 15:42:48 2022 UTC).

Diagnostics

Succeeded: s/GH rojects/GH projects/

Succeeded: s/cautions/cautious/

Succeeded: s/handle in an issue. Possibly/handle Editor assignments in an issue. Possibly/

All speakers: ktk, pchampin, pfps, souri, TallTed

Active on IRC: AndyS, AZ, Doerthe, Dominik_T, enrico, ghurlbot, gtw, ktk, olaf, pchampin, pfps, Souri, TallTed