W3C

– DRAFT –
Linked Web Storage

19 January 2026

Attendees

Present
acoburn, AZ, dmitriz, eBremer, elf-pavlik, ericP, gibsonf1, Luke, pchampin, ryey
Regrets
-
Chair
-
Scribe
AZ, pchampin

Meeting minutes

Introductions & Announcements

ericP: new introductions

Luke: I am new and I am from IBM

CRUD & Metadata PR #37

ericP: we have a PR that has been here for a while

eBremer: this has been there for a while but people asked for more time on it
… we could swap some sections and other edits
… and then other things could be done with other PRs

pchampin: I think reversing sec. 7 & 8 would be an improvement
… I have a few remarks but don't want to stand in the way so let us go ahead
… it would have been better to split this PR into pieces
… so we could discuss things separately and it would have been easier
… in the future, let us work on smaller PRs

<ericP> PROPOSED: merge PR#37 after eBremer has swapped sections 7 and 8

<gb> Issue 37 not found

<ericP> PROPOSED: merge PR#37 after eBremer has swapped sections 7 and 8

<gb> Issue 37 not found

<pchampin> w3c/lws-protocol#37

<gb> Pull Request 37 Initial CRUD with proposed metadata handling (by ebremer)

<AZ> +0 (I don't know enough about the PR)

<eBremer> +1

<acoburn> +1

<Luke> +1

<gibsonf1> +1

<ericP> +1

<pchampin> +0.5

<ryey> +1

<ericP> APPROVED: merge PR#37 after eBremer has swapped sections 7 and 8

eBremer: I'll do the changes tomorrow and merge

Storage Description Resource PR #53

acoburn: it is about describing capabilities of storages, and endpoints that extend capabilities
… discovery mechanisms fit into this PR
… we can go ahead and vote on accepting the PR

<Zakim> gibsonf, you wanted to ask about uri

gibsonf1: there is the issue of semantic URI for the storage itself different from the root container of the storage

acoburn: I think it is an important issue but it is orthogonal

pchampin: agree it is orthogonal
… with caveat that the example uses certain URIs and we need to make them better to avoid confusion

acoburn: I can record an issue for the identification issue of storage vs root container

ACTION: acoburn to add root-vs-subject issue to text merged from #53

<gb> Created action #54

<ericP> PROPOSED: merge PR#53 - Storage Description Resource

<gb> Issue 53 not found

<pchampin> +1

<eBremer> +1

<gibsonf1> +1

<acoburn> +1

<ericP> +1

<AZ> +0.5

<Luke> +1

<ericP> APPROVED: merge PR#53 - Storage Description Resource

<ryey> +0.5

"End User Credential" terminology

acoburn: will work on it this week

[acoburn shows the draft for LWS Protocol on his screen]

<gibsonf1> Agent?

acoburn: I would like to propose changing the terminology on user credential

acoburn: I'd like to call it "Authentication credential"
… I thought about alternatives: "Agent credential" but may evoke AI Agents

<eBremer> +1 "Authentication Credential"

<gibsonf1> +1 on "Authentication Credential"

<pchampin> +1

acoburn: if there is further discussion, we can have it in the PR request

Other open PRs

acoburn: the 2 issues that begin with "consolidate" related auth and ident have been open for a while
… they are more editorial than the previous 2 we takled about
… the last one is about terminology
… to clarify what we mean when we say URI
… if you have objections, please comment otherwise it will be merged by the editors

<Zakim> elf-pavlik, you wanted to ask about EUC naming

<elf-pavlik> https://www.ietf.org/archive/id/draft-ietf-oauth-identity-assertion-authz-grant-01.html#name-identity-assertion-jwt-auth

elf-pavlik: in the doc I provide above, there is something about identity and authorization, do we have the same?

acoburn: the intention is that something like this would be supported

acoburn: we could say this is identity assertion (where we say "end-suer credential" in hte draft)

elf-pavlik: [something related to the webid group]

dmitriz: the term "identify assertion" would prevent us from using zcap

acoburn: going back to the credential discussion of previous PRs
… we have different options
… we could go for identity assertion
… but could include other things

<gibsonf1> +1 on including more detail in discussion of "Authentication Credential"

elf-pavlik: we use "token" and we should be cautious
… if it's focused on authentication, then authentication credition is good

<eBremer> +1 to more detail

acoburn: I will use authentical credential and I will add more details

ericP: back to the pending PRs

elf-pavlik: related to PR31, we have a planning roadmap for SEG(?)
… this is related to access control policiies
… there are different approaches taken one related to delegation
… there is a use case related to autonoumous groups

<Zakim> gibsonf, you wanted to ask about Membership with WAC

gibsonf1: the issue that CSS(?) has is not being able to tell access
… on twinpot(?) we do ??? [something related to access control on triples]

dmitriz: a confusion is coming up often and we need to fix it

gibsonf1: single triple has its own URI, metadata, etc
… if you make a group private, what happens?

<elf-pavlik> this conversation is following up discussion about this PR w3c/lws-protocol#31

<gb> Pull Request 31 prior-art: group based access policies (by elf-pavlik)

dmitriz: if you make a URI private (ie you don't give right to access the URI)

<elf-pavlik> can't perform LWS Read operation, would that be clear?

gibsonf1: server delivers data in funciton of access right, but it must access the data to know about access rights

dmitriz: there are two separate servers, for data and authorization

[acoburn drawing a diagram on screen shared to explain the issue]

gibsonf1: how can you separate these things? how would this work?

acoburn: the issue is, there are two server models, a 1-server model and a 2-server model

<Zakim> elf-pavlik, you wanted to ask about twinpod and to mention MANDAT requirements

acoburn: a 2-server model can be specificied but it must implementable

elf-pavlik: there are cases in prior work where there are different needs on disclosing info
… in the CG group (fedid?) we plan to do a workshop to discuss cases

<elf-pavlik> Solid CG

gibsonf1: how would you do search in that scenario?
… ie if there is a different server for auth
… it would be very slow

<elf-pavlik> w3c-cg/solid#60

<gb> Issue 60 [Proposed Work Item]: Organization, Group, Team, Role support in Solid (by elf-pavlik)

gibsonf1: who has a 2-server model?

acoburn: in ACP we chose to have an ACP matcher act as a group
… but it enforces all the ACP and auth rules to be in one place

<Zakim> elf-pavlik, you wanted to discuss SAI teams/roles

elf-pavlik: depending if you have something where you can use delegation

<ericP> ADJOURNED

<elf-pavlik> use case w3c/lws-ucs#104

<gb> Issue 104 [UC] Access delegation by autonomous groups/organizations (by elf-pavlik) [triage] [usecase]

<gb> Pull Request 37 Initial CRUD with proposed metadata handling (by ebremer)

Summary of action items

  1. acoburn to add root-vs-subject issue to text merged from #53
Minutes manually created (not a transcript), formatted by scribe.perl version 248 (Mon Oct 27 20:04:16 2025 UTC).

Diagnostics

Succeeded: s/I am from IBM/I am new and I am from IBM

Succeeded: s/for this/for the identification issue of storage vs root container

Succeeded: s/liek/like

Succeeded: s/"Identification credential"/"Authentication Credential"

Succeeded: s/"Identification credential"/"Authentication credential"

Succeeded: s/dmitriz: the question is...//

Succeeded: s/ot/it

All speakers: acoburn, dmitriz, eBremer, elf-pavlik, ericP, gibsonf1, Luke, pchampin

Active on IRC: acoburn, AZ, eBremer, elf-pavlik, ericP, gibsonf1, Luke, pchampin, ryey