W3C

– DRAFT –
Linked Web Storage

07 July 2025

Attendees

Present
acoburn, bendm, eBremer, ericP, gibsonf1, hadrian, jeswr, kaefer3000, Monsecom, pchampin, RazaN, ryey, TallTed, uvdsl
Regrets
-
Chair
ericP
Scribe
kaefer3000, ericP

Meeting minutes

Introduction and announcements

ericP: Are there new introductions?

bengo: is new, was working on IPFS, is a Solid fan
… has been working with dimitriz on didcoop
… experienced from the SocialWeb WG

jsalvachua: professor, has been working in formal methods, multimedia, data spaces protocols, cloud, blockchain

<ryey> scribe?

pchampin: reports from a Geneva event
… jeswr organised a session on Solid and LWS
… the event brought a few people to LWS

<pchampin> https://globaldigitalcollaboration.org/

<bengo> https://globaldigitalcollaboration.org/agenda?day=2025-07-02

jeswr: session on Solid + LWS - discussion on whether solid can be storage for web-based credentials

<jeswr> Link to slides from Global Digital Collaboration session on LWS: https://docs.google.com/presentation/d/1nxeDF0BOv_-UyRreaL2gVPKlXVc_R5jJq3wq2AcWrV0/edit?usp=sharing

<jeswr> Link to hackMD from Global Digital Collaboration https://hackmd.io/oUdcr40QRN-SGzUWVObbzw

Use cases & requirements update

eBremer: PRs on user stories have been merged, there are PRs being reviewed, feel free to provide comments

pchampin: First draft of use-cases document published
… auto-publication system fixed
… there are HTML validation issues to be fixed
… such that merged PRs get automatically published as new version
… on w3.org/TR/

<bengo> Could someone link to FPWD of Use Cases? ty

w3c/lws-ucs#170

acoburn: eBremer opened PR about this, can we agree on a timeframe?

eBremer: plan: to be merged by the end of the week

ericP: anyone who says this is unrealistic?

TallTed: this is plenty of time

+ rationale for out-of-scope and deprioritized requirements

ericP: it's my expectation to have UCs that include requirements that are out-of-scope or de-prioritised

<acoburn> bengo: current working draft of use cases doc at https://w3c.github.io/lws-ucs/spec/. Once Echidna is fixed it will be at https://www.w3.org/TR/lws-ucs/

<bendm> +1, diff LGTM

<bengo> acoburn tysm

ericP: reasons: 1) there may be upcoming technologies that provide, 2) UCs that may inform development of upcoming technologies, 3) UCs that help non-tech people build an understanding, 4) avoidance of duplicate UCs

uvdsl: how do we determine which use-case is in or out of scope?

ericP: the decision is editorial, there will be annotations about which UC is deemed out-of-scope

acoburn: at this stage of the process, we want to be comprehensive. Some UCs will have requirements that are out-of-scope.
… imagine a UC has 5 requirements. 2 in-scope, 3 out-of-scope.
… if the 2 requirements become too burdensome, we could drop the UC entirely

uvdsl: how is the process exactly?
… when is the time to go through the list, when will be annotated?

ericP: there is no specific timeline
… during prioritisation

ericP: if you want to do a pass, that's possible

Research topics

jeswr: email on topics sent on the mailing list
… jackson and ben have answered

jackson: will work on it after Tuesday

bendm: his lab has been worked on authorisation and consent
… they could contribute

jeswr: at the moment it's about pointing to the state-of-the-art
… consent + ODRL - please do not invest too much time here, as it will be likely out-of-scope

joaquin: is working on ODRL for data spaces
… could provide review

rui: usage control is probably too far
… access control - should we collect practices closely to the Solid stack (WAC, ...) or also things farther away?
… there are other paradigms of access control

jeswr: the UC document will define what is in scope; there are input documents in the charter
… we should also be looking at how GDrive and AWS S3 are doing it
… not just Solid prior art

rui: most services use role-based access control, there is also attribute-based access control - should they get included?

jeswr: we should not doing it academically here
… if there is a nice categorisation into RBAC, ... fine

<jsalvachua> we are considering Relation based (going from Google Zanzibar and Amazon Cedar)

<jeswr> https://github.com/w3c/lws-protocol/wiki/Research-Topic-Template-(using-Authorisation-as-an-example)

<jsalvachua> and joining it to an Odrl Profile

<jsalvachua> ReBac

rui: Mastodon has access control personal, site-only, global

jeswr: provide pointers, don't make the wiki massive

<jsalvachua> we are using OpenFGA as engine for policy evaluation

<jsalvachua> I will check how it match here

bendm: what's the process? Given it's a wiki, should pointers be sent via email to the page lead who fits it in?

jeswr: put them straight into the wiki

<bengo> wrt Mastodon, it is completely different than what is required of an ActivityPub Conformant Server because it does not implement AP C2S

rui: in the wiki, not everyone has access
… how to get access?

ericP: you should be able to login
… join a help channel on IRC: #sysreq

<dmitriz> re wiki: haha, 'group membership gives access to resource' use case is hard! :)

kaefer3000: why "Research Topics"?

jeswr: i'll rename to "Prior Art"

acoburn: was looking at the GitHub configuration. Edititing is restricted to people with push access, ie. editors and chairs
… if we un-click that button, we don't know if ordinary GitHub users can edit the wiki

ericP: let's unclick and if we get undesired content, revert
… please test it

<bendm> I can edit! :)

ericP: we are through the agenda, is there any other business?

ericP: closes the meeting

Minutes manually created (not a transcript), formatted by scribe.perl version 244 (Thu Feb 27 01:23:09 2025 UTC).

Diagnostics

Succeeded: s/didcoop/on didcoop/

Succeeded: s/joaquin/jsalvachua/

Maybe present: bengo, jackson, joaquin, jsalvachua, rui

All speakers: acoburn, bendm, bengo, eBremer, ericP, jackson, jeswr, joaquin, jsalvachua, kaefer3000, pchampin, rui, TallTed, uvdsl

Active on IRC: acoburn, bendm, bengo, dmitriz, eBremer, ericP, ericp, gibsonf1, hadrian, jeswr, jsalvachua, kaefer3000, Monsecom, pchampin, RazaN, ryey, TallTed, uvdsl