[Openid-dcp] Initial results: Interop today on OpenID4VP/VCI +HAIP 1.0

[Gail Hodges]

Hello DCP WG

Below is a quick interim update on the 90min interop today on HAIP 1.0 with OpenID4VP/VCI. This is a strong start to proving out the HAIP 1.0 spec and tests. It is also the 10th in the series of interops this year.

Initial Results are encouraging:
OpenID4VP+HAIP for three DCQL queries on SD-JWT, mdoc, and both With DC API

  *
 Critical mass of 29 pairs tested so far
  *
 23 (79.3%) passing
  *
 6 with resolvable issues (20.7%)
  *
0  unknown issues.

OpenID4VP+HAIP for four DCQL queries and mdoc, SD-JWT . without DC API.

  *
6 total pairs so far
  *
3 passed (50%)
  *
2 were resolvable (33%)
  *
1 was an unidentified issue (16.7%).

OpenID4VCI +HAIP 1.0

  *
9 pairs
  *
6  passed (66.7%)
  *
3 had resolvable issues (33%)
  *
0 pairs with unknown issues

Key messages:

  *
Good results so far to prove out the specs with robust pairs on OpenID4VP side, and fewer pairs on OpenID4VCI side.
  *
2 more teams plan to join between now and end of 11/12 (Google Wallet and Ping) making a total of 8 teams including the OIDF test suite as a team
  *
Teams testing today had a good match and spectrum of functionality to allow for a good spectrum of tests between DC API /no DC-API, credential types of SD-JWT and mdoc , and a mix of DCQL queries .
  *
1 item of spec feedback , which Peter Sorotkin is requested to raise as an issue for WG clarification (note below). This was noted in DCP WG call today.
  *   1 known gap on OID4VCI test suites that was not ready yet on  key attestation, so tests were conducted and tracked as being without that feature for the wallet side. Before the next interop on Nov 12 , the Certification team is investigating if that gap can be closed to do a round of tests with that feature enabled.

  *
1 aspect of HAIP spec under review by interop participants…whether  key challenge on wallet side is  a requirement vs highly recommended, and if only highly recommended can both the implementers and tests still be updated to allow for that feature to be tested out so we can reflect it in our interop results.
  *
Oliver/Gail discussed the DCQL queries that would help offer a smooth transition to the NZTA interop 11/16-18, and we concluded that adding in more DCQL queries for 11/12 interop would be helpful. In addition, having the mdoc+SD-JWT queries enabled will further add value to the ISO interop which is not testing that configuration with two credential types. Then for the ISO results we submit, having the pair results and the SD-JWT results will assist with WG10 deliberations.
  *
Teams will continue to test asynchronously between now and the Nov 12 interop to close known gaps/ unknown gaps…and run additional pairs.
  *
OIDF plans to publish results in a blog post after the 11/12 event and the known issues are resolved, e.g. no later than end of day 11/14. We intend to include quotes from participants along with entity names/logos for those willing to have their entity name used. Gail with Serj/Elizabeth.
  *
Liaison partner (Bjorn) and Joseph (in person in NZ) can ensure results are submitted to NZTA and ISO partners.
  *
If any team not already confirmed to participate has an implementation and wants to take part — but has not indicated their intent to join—  please let gail at oidf.org know so we can add you to the group.


Gail



With DC API  - OpenID4VP+HAIP for three DCQL queries on SD-JWT, mdoc, and both. Critical mass of 29 pairs tested so far of which 23 (79.3%) passing, and 6 with resolvable issues (20.7%) and no unknown issues.
[Screenshot2025_11_06_084807.jpg]

Without DC API for OpenID4VP+HAIP for four DCQL queries we had lower number of pairs at 6 total pairs, of which 3 passed (50%) and 2 were resolvable (33%) and 1 was an unidentified issue (16.7%).
[Screenshot2025_11_06_084817.jpg]
For OpenID4VCI +HAIP 1.0 we had 9 pairs of which  6 (66.7%) passed and 3 had resolvable issues (33%)
[image.png]


Quick comment on issue to raise in Github, from Peter:

What should be used for the signing algorithms @Peter Sorotokin<https://oidf.slack.com/team/U08NT9D3WM8> can you elaborate?
We have this in the exampl for mdoc credential metadata in openid4vci: “credential_signing_alg_values_supported”: [ -7, -9 ],
8:35<https://oidf.slack.com/archives/C062U8QMVUN/p1762446934348799>
The question is: are numbers required as alg identifiers for mdoc or alg names (strings) are also must be accepted by the wallet

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs-digital-credentials-protocols/attachments/20251106/04e2906a/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screenshot2025_11_06_084807.jpg
Type: image/jpeg
Size: 221011 bytes
Desc: Screenshot2025_11_06_084807.jpg
URL: <http://lists.openid.net/pipermail/openid-specs-digital-credentials-protocols/attachments/20251106/04e2906a/attachment-0002.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screenshot2025_11_06_084817.jpg
Type: image/jpeg
Size: 211481 bytes
Desc: Screenshot2025_11_06_084817.jpg
URL: <http://lists.openid.net/pipermail/openid-specs-digital-credentials-protocols/attachments/20251106/04e2906a/attachment-0003.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 58106 bytes
Desc: image.png
URL: <http://lists.openid.net/pipermail/openid-specs-digital-credentials-protocols/attachments/20251106/04e2906a/attachment-0001.png>