W3C

– DRAFT –
Linked Web Storage

09 March 2026

Attendees

Present
!, acoburn, AZ, bendm, dmitriz, eBremer, gibsonf, gibsonf1, jessew, jeswr, laurens, Luke, pchampin, RazaN, ryey, TallTed, uvdsl
Regrets
-
Chair
acoburn
Scribe
laurens

Meeting minutes

<gb> Issue 16 Agenda: 2026-03-10 (by simoneonofri)

<pchampin> yes, 2 more weeks; back to "normal" on 30 March (at least for us in Europe)

Introduction and announcements

acoburn: Any announcements? Or changes in affiliation?

jeswr: We have a face-to-face meeting taking place on April 27th and 28th at the ODI Offices near Kings Place in London.
… We also have the Solid symposium on April 30th and May 1st.
… Finally, there is also a Solid hackathon that week.

acoburn: Jesse is encouraging everyone in the community to block out April 27th through May 1st.

Issue triage

acoburn: There is one new issue to discuss.
… Issue #95 in the lws-protocol repository.

<gb> Issue 95 Specify Container representation to be compound representation (by damooo)

acoburn: It asks a container to be specified as both a sever-managed and user-managed resource.
… I believe from the F2F that containers would be server managed data.
… I don't think we were going to specify it as being compound.
… Is #95 out of-scope for LWS or not?

gibsonf1: If a client makes an RDF request on a container would they receive everything?

acoburn: Right now containers are defined as including information on all contained resources.

<jeswr> LWS in-person meeting (Location ODI Offices, Kings Place, Kings Cross, London): April 27 - April 28

<jeswr> ODI Hackathon (pre-registration at https://theodi.org/news-and-events/news/announcing-the-solid-symposium-2026/): April 27 (afternoon) - April 29

<jeswr> Solid Symposium (https://sosy2026.eu/): April 30 - May 1

acoburn: Let's say in addition you want to add a name and other data to the container.
… At present that would not go into the container resource itself but into an auxiliary resource.

gibsonf1: So a standard GET on a container will not return basics about that resource?

acoburn: Think of a container like a folder in a filesystem.
… When you perform a GET you receive a list in the body of the resources contained in that container.
… That is what we define today as part of the representation.

gibsonf1: That is a real problem.
… If I perform a GET on a container, that returns a JSON with the content of the container.

acoburn: Currently you get a JSON response with the container contents and minimal metadata.
… Additionally there is a linkset resource associated with the container.
… Which can contain user defined and server defined link relations.
… To e.g. auxiliary resources. So the linkset allows for arbitrary (with some constraints) links to other resources.

gibsonf1: This breaks the Solid approach to getting RDF on a resource GET.
… In our case a container could be a system that contains other systems.

acoburn: Currently we are only triaging this issue. So we should probably assign a "needs-discussion" label here.

gibsonf1: There is an Accept that you send on the GET request.

acoburn: That is part of HTTP.
… We require that a server support a specific mediatype (application/lws+json).
… A server may do content negotiation on that resource.

Timeframe for FPWD

acoburn: Typically specifications start with an editor's draft.
… And after some time you move into a first public working draft.
… It is a signal that (while still being a draft) it is a publication which begins a process
… of patent exclusion
… If there is IP in the draft exclusion requests can be made. This starts at the FPWD.
… This process takes a couple of months.
… Our charter expires in September, so we want to finish this patent exclusion process before that.
… We are probably going to need an extension anyway, but it would be good to move along to FPWD.

pchampin: A first public working draft is still a draft.
… We shouldn't be shy or wait for the content of the specification to be more mature.
… Some groups publish an empty FPWD.
… Just to have a stake in the ground.
… We still need to improve a lot in the spec, but we should release early and often.
… It does not yet represent consensus of the WG other than consensus that we should put something out there.
… Our editor's draft is already visible today.
… But it does start this patent exclusion call.

acoburn: I have two resolutions ready, one on the core specification and one on the authentication suites.

<acoburn> PROPOSED: The LWS WG shall publish the FPWD of the core LWS specification

<laurens> +1

<Luke> +1

<pchampin> +1

<eBremer> +1

<acoburn> +1

<TallTed> +1

<dmitriz> +1

<ryey> +1

<gibsonf1> +1 (with small change to pagination pr)

<gibsonf1> Ahh, sorry +1

acoburn: It looks like we have solid consensus

RESOLUTION: The LWS WG shall publish the FPWD of the core LWS specification

<gb> Issue 16 Agenda: 2026-03-10 (by simoneonofri)

<acoburn> PROPOSED: The LWS WG shall publish the various authentication suites as FPWD documents

<gibsonf1> +1

<laurens> +1

<Luke> +1

<acoburn> +1

<pchampin> +1

<TallTed> +1

<bendm> +1

<acoburn> Core specification: https://w3c.github.io/lws-protocol/lws10-core/

<acoburn> OpenID: https://w3c.github.io/lws-protocol/lws10-authn-openid/

<acoburn> SAML: https://w3c.github.io/lws-protocol/lws10-authn-saml/

<acoburn> SSI-CID: https://w3c.github.io/lws-protocol/lws10-authn-ssi-cid

<acoburn> SSI-DID-KEY: https://w3c.github.io/lws-protocol/lws10-authn-ssi-did-key

<ryey> +1 (after fixing some seeming editorial issues?)

acoburn: We also have consensus for the authentication suites.

<eBremer> +!

<eBremer> +1

<acoburn> PROPOSED: The LWS WG shall publish the various authentication suites as FPWD documents: https://w3c.github.io/lws-protocol/lws10-authn-openid/, https://w3c.github.io/lws-protocol/lws10-authn-saml/, https://w3c.github.io/lws-protocol/lws10-authn-ssi-cid, https://w3c.github.io/lws-protocol/lws10-authn-ssi-did-key

RESOLUTION: The LWS WG shall publish the various authentication suites as FPWD documents: https://w3c.github.io/lws-protocol/lws10-authn-openid/, https://w3c.github.io/lws-protocol/lws10-authn-saml/, https://w3c.github.io/lws-protocol/lws10-authn-ssi-cid, https://w3c.github.io/lws-protocol/lws10-authn-ssi-did-key

Remaining priorities for WG: test suite, access requests, notifications, type index

acoburn: We have about 6 months left in the charter.
… We are very much interested in renewing the charter.
… We need to set priorities for both the next six months and think about what comes after.
… Probably after those six months the focus will be on implementation.
… But by September the contents of LWS should be mostly complete.
… We have a couple of things that haven't been taken up: test suite, access requests & grants, notifications, server-managed type index.
… Some other issues still need attention like pagination, multiple containment, ...
… But the foundation is already there for containment.
… However on these topics we haven't started working yet.
… Are these still the right priorities? Are we missing anything?

<gibsonf1> +1 on type search (and I can write a proposal up for it)

ryey: For test suites, do we test the server implementation? Or also the client?

acoburn: We are primarily specifying a protocol.
… Any entity for which we're specifying behaviour we would want to test.

<dmitriz> do we have a use case defined for type index?

acoburn: We're primarily focusing on server behaviour.
… So that will be our focus in testing. And will give the greatest benefit.
… But we could also have tests for client behaviour. But that might be more complicated.

<TallTed> "conformant by assertion"

acoburn: That's the test suite.
… I see gibsonf1 giving a +1 on type index.
… The main use case is data discovery.
… I don't have a specific reference, but we do have cases for that.

<Zakim> bendm, you wanted to state access request & grant is authorization server, and may be too far for the timeline, notifications seems manageable in the timeline, for server-managed type index I would suggest an extensible system (and happy to read gibsonf1's suggestion)

bendm: Test suite is something we must do.
… Access request & grant is related to authz service so farther from storage and thus less of a priority.
… Notifications seems manageable in the timeline.
… A server-managed type index will be hard to agree upon.
… So an extensible system will be needed.

gibsonf1: On the type index, I'm not thinking of it as a way of indexing things
… but as a search on types.
… A request would be one or more types which can be intersected.

acoburn: That would be great. As soon as we have a proposal we can discuss further.
… We will have to define the scope and requirements as a group.
… Please let one of the chairs know as soon as you have something ready.
… With respect to access requests & grants Inrupt is very interested.
… I could probably bring something to the group by next week.

<Zakim> ryey, you wanted to ask about "type" in type index and lack of RDF-native support

laurens: Access Requests & grants is very important still and resolves a severe lack of the solid protoocl.

ryey: I would agree on the access requests & grants.
… We are talking about type indexes, but as we've dropped the RDF support
… What do we mean by the type of a resource?

acoburn: This would probably be part of the scope of type indexes.
… In an LWS storage there will be both RDF and non-RDF resources.

How can we support as broad a category as possible.
… Types could be represented as part of the linkset.

<Zakim> bendm, you wanted to say we're also very interested in access request and grants, we can also bring something to collaborate on, I'm just wondering how large of a specification that will add to v1.0

bendm: Access requests & grants are super important.
… But it will be a new and big thing to add to the specification.
… So the question is how far we want to go.
… I might be wrong and it might be more straightforward.
… I just wonder if it wouldn't be better to scope ourselves.

acoburn: I have tried to keep this as minimal as possible.
… So the goal for access requests & grants would be to be as lightweight as we can.

bendm: I do think notifications are very important, as it allows for extensibility.

luke: The access requests & grants seems to be one of the most user unfriendly things that is in place right now.
… So it is important in terms of adoption to add this.

acoburn: I'm not hearing anything new.
… Only concerns on access requests & grants and their complexity.
… We need to make sure there are individuals in the WG actively working on these items.

<uvdsl> (sorry for joining just before the end :) )

acoburn: If anyone else wants to collaborate on access requests, let me know.
… gibsonf1 has indicated interest in working on type indexes. If anyone else is interested, get in touch.
… That leaves test suite and notifications.

<Zakim> dmitriz, you wanted to comment on test suite

acoburn: I am happy to contribute to notifications, but would be happy to have some assistance.

dmitriz: The test suite is a special case
… The W3C policy tends to strongly encourage multiple interoperable implementations.
… So it is an existential item to get anything out of the door.

jeswr: There is a grant from NLNet
… That has been granted to ODI for a test suite.
… That will be worked on.

acoburn: It would be wonderful if we have some collaboration on writing that test suite.

gibsonf1: There is a solid test suite group currently active
… Wouldn't it make sense to bring them in.

acoburn: They have a test suite written using Karate
… We could build on there work, so that would be great.

Web-CID Authentication Suite

acoburn: This PR proposes another authentication suite.
… uvdsl could you provide an introduction to this PR.
… So that we can discuss and vote on it next week.

uvdsl: This proposal is currently on agent identification, not authentication.
… As identification is a fundamental component of authentication.

<uvdsl> w3c/lws-protocol#57

<gb> https://github.com/w3c/lws-protocol/issues/57

uvdsl: Following #57 I was tasked to define a conformance profile on agent identification.

<gb> Issue 57 Agent Identification (by uvdsl) [ready-for-pr]

<uvdsl> https://uvdsl.solid.aifb.kit.edu/spec/lws/ident-cid-web/draft.html

uvdsl: My approach is to define the mechanism for dereferencing a CID by a client.
… The specification should be implementable and lean.
… Such that it can be used in the existing authentication suites.

acoburn: I encourage everyone to look at this PR.
… With that we are out of time.
… Any last thoughts?

Summary of resolutions

  1. The LWS WG shall publish the FPWD of the core LWS specification
  2. The LWS WG shall publish the various authentication suites as FPWD documents: https://w3c.github.io/lws-protocol/lws10-authn-openid/, https://w3c.github.io/lws-protocol/lws10-authn-saml/, https://w3c.github.io/lws-protocol/lws10-authn-ssi-cid, https://w3c.github.io/lws-protocol/lws10-authn-ssi-did-key
Minutes manually created (not a transcript), formatted by scribe.perl version 248 (Mon Oct 27 20:04:16 2025 UTC).

Diagnostics

All speakers: acoburn, bendm, dmitriz, gibsonf1, jeswr, laurens, luke, pchampin, ryey, uvdsl

Active on IRC: acoburn, bendm, dmitriz, eBremer, gibsonf1, jeswr, laurens, Luke, pchampin, ryey, TallTed, uvdsl