Protocols and APIs should remain out of scope

[selfissued]

The charter at https://www.w3.org/2017/vc/WG/charter.html states that APIs and protocols for using the Verifiable Credentials Data Model is out of scope. Yet the new charter adds possible non-normative deliverables such as "Storage and Sharing of Verifiable Credentials" and "An API for Verifiable Credential Exchange", which appear to creep into that territory. The working group should focus on creating a practical and useful data model, while leaving choices of platform APIs and use in communication protocols to those building and deploying systems using the data model.

Please explicitly state that protocol and API work are out of scope and remove possible deliverables that infringe on these areas.

[OR13]

I agree the data model is the first priority, but I question if keeping protocols out of scope will prevent real interoperability...

Regarding credential exchange protocols, we have a growing ecosystem of mostly not interoperable protocols... A short list of them:

• DIDComm v2
• OIDC
• OIDC Presentation Exchange?
• WACI Presentation Exchange
• W3C CCG VC API
• Credential Handler API

Obviously we can't work on OIDC in the W3C, but not being able to talk about it might make interop harder to achieve.

And the charter will be trivial to attack if protocol interop is out of scope:

"See, you don't have enough interop, you just defined data models, you need protocols for practical interop"

We saw this with DID Core... a JSON data model was not enough, we needed to define resolution better.

I share the concern regarding charter scope creep, but I am also concerned about fixing the charter scope too narrowly, and then seeing formal objections or "near objections" regarding interoperability, when our charter may not allow us to address those concerns.

I think we need to be able to talk about protocols in order to do a good job, but I am not sure we need to do so normatively.

[iherman]

The issue was discussed in a meeting on 2021-12-01

• no resolutions were taken

View the transcript

1.4. Protocols and APIs should remain out of scope (issue vc-wg-charter#24)

See github issue vc-wg-charter#24.

Brent Zundel: Mike has expressed his concerns, Orie has responded, anyone else want to comment?.

Manu Sporny: we're not saying we're gonna do normative things here.
… but we are absolutely incubating a VC API spec.
… if one of the things we want to do is demonstrate an end-to-end ecosystem, the VC API could get us there pretty quickly.
… multiple people are implementing, we're demonstrating interop around mDL, etc..
… it's a useful thing to bring into the conversation, without saying we're going to standardize it.
… so that's the reason to have it in scope -- the work /is/ happening..
… and we'll get a lot of questions "how do you know this VC data model works, interops"? And one of our answers could be -- well, we have all these interop tests.

Phil Archer: Trying to look into Mike's comment -- 2017 charter says things are out of scope... we're talking about the new charter... it was out of scope, but it's no longer out of scope, don't see what the problem is..

[iherman]

The issue was discussed in a meeting on 2021-12-08

• no resolutions were taken

View the transcript

1.3. Protocols and APIs should remain out of scope (issue vc-wg-charter#24)

See github issue vc-wg-charter#24.

Brent Zundel: We talked about this issue during our last meeting, Orie made some comments, still no response from Mike Jones. Any recommendations on how to move this issue forward?.

Manu Sporny: I think we should close the issue. People want to talk about this, work is proceeding, Mike needs to make the case for why the non-normative version of this needs to be out of scope, why we can't talk about it at all. It's a strange ask..

Manu Sporny: +1 to pending close, we need a response..

Brent Zundel: We could put a pending close label on it and ping Mike Jones and see if he can respond to the feedback and make his case. Any opposition to that?.

Dave Longley: +1 to that approach..

Brent Zundel: If we still haven't heard from Mike the next time we meet we can close it..

[brentzundel]

no response since marked pending close, closing.

[iherman]

The issue was discussed in a meeting on 2021-12-15

• no resolutions were taken

View the transcript

1.1. Protocols and APIs should remain out of scope (issue vc-wg-charter#24)

See github issue vc-wg-charter#24.

Brent Zundel: first issue is number 24.
… last time we said we'd close it so I move to close the issue any opposed?.
… none heard going to close this one.

David Chadwick: I did tell Mike last week about this.

[brentzundel]

Reopening at request of @selfissued

[OR13]

I think we should move identifier -> publicKey in scope, otherwise everyone will invent their own way of reinventing dereferencing documentLoaders

[Sakurann]

Do you mean mandating identifiers in VCs to be tied to a public key in some manner? ie identifier being referenceable to a pubKey or being a PubKey itself?

[iherman]

I think we should move identifier -> publicKey in scope, otherwise everyone will invent their own way of reinventing dereferencing documentLoaders

@OR13 is this related to the current issue (API-s) or is it a comment on the spec itself (e.g., §4.2)? Sorry if I miss something obvious.