Replace "Client ID Scheme" with "Client ID Prefix"

[aaronpk]

This is to help avoid the confusion encountered in #376 by avoiding the use of the term "scheme" which implies that the client ID value is a URL. Using the term "prefix" avoids this association while still being descriptive.

[selfissued]

"Prefix" is better.

[c2bo]

Agreed, prefix is probably a better term.

I just did a quick glance through the changes and noticed one part that needs to be adjusted: line 2239

The client_id parameter MUST be omitted in unsigned requests defined in (#unsigned_request). The Wallet determines the effective Client Identifier from the Origin. The effective Client Identifier is composed of a synthetic Client Identifier Scheme of web-origin and the Origin itself. For example, an Origin of https://verifier.example.com would result in an effective Client Identifier of web-origin:https://verifier.example.com. The transport of the request and Origin to the Wallet is platform-specific and is out of scope of OpenID4VP over the W3C Digital Credentials API. The Wallet MUST ignore any client_id parameter that is present in an unsigned request.

[jogu]

Thanks for doing this Aaron, it's appreciated! Unfortunately I'm pretty sure this will conflict with #448 :( It would be good to merge that first if possible - I think it just needs one more approval before we'd be ready to merge it. (The line Christian mentions in his comment above is modified by 448 too.)

[aaronpk]

I will gladly update this PR after #448 is merged

[Sakurann]

chair hat off

[Sakurann]

chair hat on. this is a pretty big breaking change. @aaronpk can we please discuss these kind of changes/proposals in the WG before doing a PR? thank you.

[Sakurann]

thinking more about this. this might be a better name, but I want implementer feedback first before we introduce these kind of breaking changes (in the metadata and error codes, but still) for the aesthetics reasons, at this point in time.

[jogu]

@aaronpk #448 is merged now if you'd like to update this one to resolve the conflicts.

@Sakurann

thinking more about this. this might be a better name, but I want implementer feedback first before we introduce these kind of breaking changes (in the metadata and error codes, but still) for the aesthetics reasons, at this point in time.

I agree with getting implementer feedback, but I think the impact is fairly limited - wallet metadata isn't used much, and the change in the error is in the error_description field which isn't intended to be machine readable. (The error code remains invalid_request unless I missed a change somewhere.)

[aaronpk]

chair hat on. this is a pretty big breaking change. @aaronpk can we please discuss these kind of changes/proposals in the WG before doing a PR? thank you.

It seemed easier to submit the PR to start the discussion so folks have a more concrete proposal to work from. It's almost entirely an editorial change. The only actual protocol change is changing the metadata parameter from client_id_schemes_supported to client_id_prefixes_supported. The other place the term appears on the wire is in the error_description which should not be being read by machines anyway.

[aaronpk]

I updated this PR and resolved the conflicts from #448

[babisRoutis]

Prefix is a better term and changes (to implementations) are limited to the wallet metatada, used mostly - I guess - conditionally, in case of a request_uri_method equal to post.

LGTM

[Sakurann]

approving since it looks like the benefits of this naming change outweights my worrys about a small breaking change.

[Sakurann]

@c2bo to open issues in VCI and HAIP to fix terminology there

[jogu]

Merging this as agree on today's WG call - and as agreed I'll update #493 to move the changelog entry to -26.