Re: Call for Co-Editors: Post-Quantum Migration Pattern for Verifiable Credentials (Ed25519 + ML-DSA Dual-Signature Model)

> I also agree with you, Jori, that the QR spec should note how you do the
transition easily with proof sets.

Yes, this is the pragmatic approach.

> IIUC, this is already done because we foresaw the problem over a decade
ago. Data Integrity allows proof sets that are cryptographic
algorithm-agnostic. Always has.

Proof sets do provide flexibility, and the data model does seem to allow
all the transition scenarios I can think of.

I was not aware of proof sets. Even without them, issuing two separate VCs
would of course be a possible workaround.

Where I may disagree is that, with or without proof sets, we seem to need
the capability to resolve multiple Data Integrity suites instead of
resolving multiple algorithms within one suite. I doubt there is much
difference in canonicalization, hashing, and other preprocessing between
PQC and ECC suites. I guess it is possible to use mostly the same code for
both and simply choose the algorithm based on the declared suite, but it
feels counterintuitive to talk about different Data Integrity suites if the
only meaningful difference is the algorithm.

And as a side note, we have had discussions worried about signature size,
byte size, and performance. Yet here we are ready to grow the data model
and required runtime by using multiple proofs and suites.

And as another side note I'm most likely not aware of some critical
information related to the differences between the PQC and ECC suites...

Nothing critical here; these are just my personal biases I wanted to share,
and maybe I am sensing some form of a double standard ; ). But sometimes
the most important thing is simply to get the job done, and the great work
from the group does that.

I also have not been actively contributing... so these are just some
thoughts that are making me personally second guess using these standards
in my projects, and me communicating in what kind of specs/projects I would
love to contribute to.

But, here I agree that dedicating a section to documenting algorithms on
how to cleanly migrate using the existing data model is the productive and
right thing to do.


su 14.6.2026 klo 15.42 Manu Sporny (msporny@digitalbazaar.com) kirjoitti:

> On Sun, Jun 14, 2026 at 5:38 AM Jori Lehtinen <lehtinenjori03@gmail.com>
> wrote:
> > I’m not sure whether there is already a PQC Data Integrity suite on the
> W3C standards track.
>
> There is:
>
> https://w3c.github.io/vc-di-quantum-resistant/
>
> It will be published as a first public working draft this coming week
> (and it is standards-track at W3C):
>
> https://w3c.github.io/vc-wg/minutes/2026-05-27.html#5aec
> https://github.com/w3c/transitions/issues/806#issuecomment-4575832272
>
> > That experimental draft is what I found quickly, and perhaps a section
> on hybrid signatures and/or pqc migration in general could be added to it.
>
> We're not going to do hybrid signatures. IMHO, hybrid is a waste of
> time given that the architecture of Data Integrity allows you to do
> proof sets. You've always been able to do that with Data Integrity
> (going on 10+ years now):
>
> https://www.w3.org/TR/vc-data-integrity-1.1/#proof-sets
>
> So, the W3C VC + Data Integrity solution is to just put a parallel
> signature on the VC. I believe Amir is saying that we should document
> that more clearly. I also agree with you, Jori, that the QR spec
> should note how you do the transition easily with proof sets.
>
> I do also acknowledge that JOSE/COSE has no such capability that is
> broadly deployed today, but that's a design decision that technology
> stack made and it's not viewed as an architectural failure there. I
> don't think the VCWG or CCG has the bandwidth to "fix" JOSE/COSE at
> this point wrt. proof sets given that it's not seen as an issue in
> that community. That's where hybrid would make more sense, but it
> really does feel like they're out of time due to the accelerated
> transition timeline.
>
> > But for me, the requirement would be that we specify an
> algorithm-agnostic key management solution and an algorithm-agnostic Data
> Integrity suite that can handle whichever algorithms are recommended over
> time.
>
> IIUC, this is already done because we foresaw the problem over a
> decade ago. Data Integrity allows proof sets that are cryptographic
> algorithm-agnostic. Always has.
>
> That said, people continue to not understand that this capability
> exists... and you can't blame them, it's buried in the global
> standards, and not many people read those.
>
> I agree with Amir that we need more people to write about this
> capability and that there is a fairly straight-forward, non-disruptive
> solution that can be deployed that can easily weather the transition
> to a post-quantum world.
>
> -- manu
>
> --
> Manu Sporny - https://www.linkedin.com/in/manusporny/
> Founder/CEO - Digital Bazaar, Inc.
> https://www.digitalbazaar.com/
>
>

Received on Sunday, 14 June 2026 16:35:15 UTC