W3C

– DRAFT –
LWS WG meeting (2025-02-17)

17 February 2025

Attendees

Present
AZ, bendm, csarven, dmitriz, eBremer, ericP, jesse, laurens, pchampin, ryey
Regrets
-
Chair
laurens
Scribe
pchampin

Meeting minutes

Introductions & Announcements

laurens: reminder: please try to keep under two minutes when you have the floot. This applies to anyone, including the chairs.
… Any other announcement?
… Hearing none, let's move on.

Action items

laurens: I see two open actions.
… Should we keep "sticky" action items open? I'd rather not.
… We should have a different way to capture recurring topics.
… Any suggestion?

ericP: should be on a case-per-case basis. Some people would prefer being pinged regularly.
… If not, we can close those actions.

laurens: I'll ask Hadrian how he prefers to deal this those.
… It is not clear what the definition of "done" is for those actions.

Template for requirements (w3c/lws-ucs#119)

laurens: There has been some discussions on this PR already.
… csarven, you made some comments. Do you want to develop?

csarven: I don't have anything to add. pchampin's suggestions on the PR are good.
… They are pending response by Hadrian. Review from others would be good.

laurens: agreed; everyone is encouraged to give feedback on this PR.

Consensus on glossary (w3c/lws-ucs#122)

laurens: there was also some discussion in this issue. We did discuss it last week.
… I would like to discuss further the choice between "owner" and "controller".
… Hadrian made a point about "controller", which GDPR uses differently.

dmitriz: great point, we could call it out explicitly ("we use this term differently from GDPR").
… The thing we call "owner" in Solid is the person/entity allowed to make changes to the resource and its metadata, hence "controller".
… The same discussion occurred in the Verifiable Credentials / Decentralized Identifiers group, where they first had "owner", the decided to rename it to "controller".

laurens: +1 to make it explicit that we do not use it the same way as GDPR.

csarven: in the use cases, some aspects are more related to "owner", others more to "controller".
… I agree with what dmitriz said.
… There is an interaction between owner and controller. There is a question of liability.

laurens: Do you consider liability as the separation line between owner and controller?

<dmitriz> I don't think any of our usecases use the ownwer notion tho..

csarven: yes, that's a way. In WebArch, the term owner is used for "URI owner".
… There is no direct mapping to something like a controller.

<ericP> dmitriz, i'll report that when i get called upon

csarven: We should be clear in the spec that "owner" may not be the same as "URI owner".

<Zakim> ericP, you wanted to ask if the controler is different from the person who has legal/adminstrative ownership and whether that matters in these use cases

csarven: I support dmitriz's suggestion, but not to the point of dropping the term "owner" completely.

ericP: to ask if the controler is different from the person who has legal/adminstrative ownership and whether that matters in these use cases

<dmitriz> thx

pchampin: what we call an "owner" in webarch is usually a renter. we need to be clear about how we differ from other specs' use of the term

ryey: people talk about differen things simultaneously. We need to make a difference between owner, controller and admin.
… one is the administrator of the storage service.

laurens: I agree that we also need to discuss the notion of administrator.
… But do we need to consider that?

laurens: I see some consensus on introducing the term "controller", but not as a full replacement of the term "owner".

<bendm> +1

<Zakim> ericP, you wanted to propose we clarfiy our set of terms and offer them to the UC&R editor to include if referenced in the document

ericP: I was goint to propose the same thing. Do we need to define them, or defer to Hadrian for that?

<laurens> PROPOSAL: To introduce both the term owner and controller in the glossary, and clarify their meaning.

<AZ> +0.8

<laurens> +1

ericP: with some caveat about "owner" often being a "renter"

<ericP>

laurens: not necessarily a resolution, but let's make an action for Hadrian

<pchampin> +0.8 (not entirely sure that two notions are required, but not opposed to it)

<bendm> +1

ACTION: hadrian To introduce both the term owner and controller in the glossary, and clarify their meaning.

<gb> Created action #15

<csarven> +1 iff the use cases cover it

laurens: any other topic to discuss for the glossary? We touched on the notion of administrator?

pchampin: Solid/LWS is not always about self-hosting, so distinguishing administrator from controller is useful.

Potential blocking items, progress & timeline for UCR document

laurens: personally, I'd like a first draft of the UC note in the coming weeks.
… I've discussed that with Hadrian asynchronously.
… I would also like to start work on the protocol soon.

<pchampin> +1 to this (rough) timeline

laurens: we need the glossary and requirements template in place.
… Is there anything that is blocking us?
… Anyone wants to add anything about the UC note, or the upcoming Protocol spec?

csarven: will we be discussing the glossary again today?

laurens: I'm not sure we can progress much without Hadrian.

csarven: other topic: we will need to demonstrate adequate implementation experience.
… Should we have a survey on what people are planning to implement?

laurens: +1 about the idea of getting commitment to implement.
… But I'm not sure how we can approach that. At least we can gauge interest in implementing given use-cases?
… We can put this discussion on the agenda for next week.

ack

Capturing interest in implementation

csarven: we don't need to hash out some classes of use-cases or requirements if no one intends to implement them.
… At the end of the day, the W3C criteria requires that we have implementation of all features.
… Discussing interest in implementation beforehand will save us some time.

laurens: +1 on the focus and prioritozation

<csarven> User stories / cases survey, e.g., in Social Web WG https://www.w3.org/wiki/Socialwg/Social_API/User_stories , Solid CG: https://github.com/solid/authorization-panel/blob/main/proposals/authorization-ucr/uc-survey.md

pchampin: deciding whether requirements/UC are in- or out of scope should definitely be based on the interest of implementers

csarven: see links above, how the Social Web WG and Solid CG captured the interest of the community
… It is good insight.
… It is also part of bookkeeping (compare actual implementations with declared intents).

laurens: good proposal. I would like to discuss it with the editors.
… Some organization may not want to share on the open what they plan to implement, but for us the transparency is good.

csarven: it is optional. These organization don't have to share. If it is useful for 80%+ of us, that's already good.
… It should also be clear that those "commitments" are not legally binding :)

laurens: I'll ping Hadrian about this discussion. Hopefully he is with us next week to continue the discussion.

AOB

jesse: there has been discussion in the CG, about whether the CG can make changes to the Solid protocol CG report.

<jeswr> https://lists.w3.org/Archives/Public/public-lws-wg/2025Feb/0000.html

jesse: I'd like a WG decision to allow the CG to pursue.

<jeswr> solid/specification#711

<gb> Pull Request 711 change: allow blank nodes to be inserted in N3 Patches (by jeswr) [doc: Protocol]

laurens: unfortunately, we need more that 3 minutes.
… I will send an email to the mailing list and put this in the agenda for next week.

<laurens> CG scope, and whether changes to the CG publications can be made./ACTION: laurens to write an e-mail to the mailing list on the WG/CG scope, and whether changes to the CG publications can be made./CG scope, and whether changes to the CG publications can be made.

<gb> Cannot create action. Validation failed. Maybe to write an e-mail is not a valid user for w3c/lws-protocol?

Summary of action items

  1. hadrian To introduce both the term owner and controller in the glossary, and clarify their meaning.
Minutes manually created (not a transcript), formatted by scribe.perl version 242 (Fri Dec 20 18:32:17 2024 UTC).

Diagnostics

Succeeded: s/Z> present+//

Succeeded: s/fisrt/first

Succeeded: s/in use cases/in implementation

Warning: ‘s/ACTION: to write an e-mail to the mailing list on the WG/CG scope, and whether changes to the CG publications can be made./ACTION: laurens to write an e-mail to the mailing list on the WG/CG scope, and whether changes to the CG publications can be made./’ interpreted as replacing ‘ACTION: to write an e-mail to the mailing list on the WG’ by ‘CG scope, and whether changes to the CG publications can be made./ACTION: laurens to write an e-mail to the mailing list on the WG/CG scope, and whether changes to the CG publications can be made.’

Succeeded: s/ACTION: to write an e-mail to the mailing list on the WG/CG scope, and whether changes to the CG publications can be made./ACTION: laurens to write an e-mail to the mailing list on the WG/CG scope, and whether changes to the CG publications can be made./

All speakers: csarven, dmitriz, ericP, jesse, laurens, pchampin, ryey

Active on IRC: AZ, bendm, csarven, dmitriz, eBremer, ericP, jeswr, laurens, pchampin, ryey