RDF-star WG

<pfps> ls

Last call's minutes

https://www.w3.org/2023/01/26-rdf-star-minutes.html

<Souri> +1

<enrico> +1

<gkellogg> +1

<ora> +1

<AZ> +1

<pchampin> +1

<Timothe> +1

RESOLUTION: minutes approved

"RDF 1.2 vs. RDF-star"

AZ: I will not repeat everything I wrote in my email
… but the gist of it was that I was suprised when the charter was proposed
… I think there could be an option to standardize RDF-star and SPARQL-star as a separate standard,
… leave some time for implementations evolve
… then standardize RDF 1.2 and SPARQL 1.2
… this would avoid detrimental impact on other standards built on top of RDF.
… Then Ora propose that RDF/SPARQL 1.2 would have two kinds of comformance :
… full (with quoted triples) and weak (no quoted triples).
… I would still prefer the 1st option, but could live with the 2nd option.

Ora: I think that this option to have 2 levels of comformance is within what the charter allows us to do.
… But we should then quickly define what the RDF-star features will be.
… Given the landscape of graph databases and labeld property graphs,
… we need RDF-star as a standard.
… Backward compatibility is really important -- unless we find a very good reason to break it.

pchampin: Content Negotiation by Profile https://www.w3.org/TR/dx-prof-conneg/

<Zakim> gkellogg, you wanted to suggest that this leads to fragmentation in the marketplace.

gkellogg: there is a slippery slope here; risk of fragmentation of the marketplace.
… Also, some implementations would more easily support RDF-star if all quoted triples were asserted.
… That's another thing to consider.
… Technically, this is possible, but let's gather some feedback before we decide to go there.

ora: I appreciate the risk of fragmentation,
… but the fragmentation may happen anyway.
… A standard is better than no standard.
… Profiles seem indeed a right way to go.

AndyS: it would work if we defined RDF 1.2 with a set of profiles for restricting it.
… It would not work if we had [RDF 1.2 + RDF-star on top of it].

Souri: by profile, do you mean the same kind of thing as OWL profiles ?

pchampin: yes, that's a good analogy IMO.

ora: but the DX work item is also considering content-negociation, which OWL profiles do not cover.
… Question to AndyS: can you exmplain what you said earlier about referencing features?

AndyS: when SPARQL 1.1 was standardise, some features were conditional to the (future) publication of RDF 1.1.

ora: we don't need to decide this now.
… Let's continue the discussion on the mailing list.

editors assignment

w3c/rdf-star-wg#11

gkellogg: glad to see so much interest in the existing specs
… the new ones (rdf-new, sparql-new ...) are typically handled at the end, usually lead by the chairs
… the sparql-concepts can start from the current sparql11-overview document
… it is generally expected to have more than 1 editor per spec, which we have on most of them

ora: happy to have my name on the *-new documents

<Zakim> gkellogg, you wanted to discuss practical matters

gkellogg: we are in the middle of an experiment of using submodules, which I am currently reverting

<gkellogg> https://github.com/w3c/rdf-star-wg/wiki/Editor's-guide

gkellogg: so wait until tomorrow until you start working on the repos.
… There is also an Editor's guide (link above).
… We had an impromptu call with AndyS and Greg Williams.

ora: can you send a mail on the mailing list when it is ok to clone?

gkellogg: will do.

<Zakim> pchampin, you wanted to ask about the editors call

gkellogg: when putting yourself as editor in the Respec document,
… you need to put your W3C ID in the JSON.
… This number happens in the URL of your W3C account page.

https://www.w3.org/users/myprofile

<gkellogg> All editors MUST have a w3cid field, other uses may include it as well.

ora: this is necessary if you are an editor; not to do pull requests

<Zakim> AndyS, you wanted to ask about non-WG PRs

gkellogg: the W3C id for editors is required by the W3C validation process.

AndyS: are we accepting PRs from outside the WG participants? If so, how does that work?

gkellogg: if a PR is made by someone external, an intervention is needed
… to ensure that the contributors comply with the patent policy.

AndyS: if an individual makes a PR, who is an employee of a member organization,
… how does it work? Has the ACrep of the organization to get involved?

ora: I guess it has to go through the ACrep.

ACTION: pchampin to check what happens with contribution from member's employees

<ghurlbot> Created action 13

gkellogg: accept external contributions on a per-case basis
… another class of contributions is the ones by the bots

pchampin: would it make sense to add the editor's ad-hoc calls in the WG calendar?
… pros: creates awareness; cons: clutters everyone's calendar

gkellogg: different expertise in Respec; we can schedule a call for some kind of "tutorial"
… but I think that most people who have signed up have some experience in Respec.
… Don't want to be the "respec dictator".

ACTION: pchampin to propose a doodle for an editors call next week

<ghurlbot> Created action 14

gkellogg: a number of changes will need the group's approval
… but to move forward, we should not require *every* PR to get the group's approval
… It is always possible to open an issue if the editor's didn't reflect the group's consensus.
… We might want to be more structured once a FPWD is published.

<Zakim> ora, you wanted to discuss next week's agenda

ora: almost out of time; anybody wants to talk about use-cases?

Use cases

enrico: I believe there is a class of use-cases, IMO the typical ones for reification,
… for which the current formalization does not work.
… Not saying that what has been proposed is wrong,
… but that we need something in addition.
… Something where reified triples are fully transparents.

<Zakim> AndyS, you wanted to suggest we have a UCR editor (whether it is formally-W3C-published or not)

pfps: I'm uncomfortable discussing technical details without concrete use-cases.
… So let's create the use-cases documents quickly.

AndyS: suggest to find editors for a UCR document, and create the document.

enrico: I appreciate your contrution to this group, it is needed.
… Also agree with pfps about use-cases.
… In favour of AndyS' suggestion.

pchampin: should I add the ucr document in w3c/rdf-star-wg#11 ?

ora: maybe mentioning that it might not get published in the end
… but still a useful document to have.
… Thanks everyone, adjourned.

<TallTed> I suggest a doc within the rdf-star-wg repo for the UCR, intended to be delivered as a WG NOTE

<pchampin> s/^ls$//

<pchampin> s|s/^ls$//||