Hi Filip, Thanks for pointing this out, I have raised a PR to clarify:
https://github.com/cose-wg/draft-ietf-cose-dilithium/pull/15/files#diff-2478aa1dfa824cbb11926553a0f4f8c9cd3e223126798f02e4827a8a56e809e9 Please let me know if you disagree with the proposed changes above ^ Regards, OS On Wed, Mar 19, 2025 at 7:08 PM Filip Skokan <[email protected]> wrote: > Hello Orie, > > I'm late to the discussion here but I'd like to point out that > the Algorithm Key Pair Type does not specify whether "pub" must always be > present when a private key is represented. > > e.g. https://www.rfc-editor.org/rfc/rfc7518.html#section-6.2.2 >> In addition to the members used to represent Elliptic Curve public keys, >> the following member MUST be present to represent Elliptic Curve private >> keys. > > > Currently, all asymmetric JWK formats are defined such that private keys > always include their public key components. For AKP that is not the case > and the document is silent on the topic. > > S pozdravem, > *Filip Skokan* > > Hi Paul, > > I've not been able to keep up with that discussion, but I believe you are > correct. > > At some point we will probably see a similar discussion for ML-KEM and > x-wing. > > It would be unfortunate if COSE/JOSE and LAMPs ended up diverging on the > bare key and signature formats for PQC. > > Regards, > > OS > > > On Tue, Mar 11, 2025, 11:33 AM Paul Wouters > <[email protected]> <[email protected]>> > wrote: > > > Hi, > > > > The document looks fine to me, but I'd like to discuss one item. > > > > It seems private keys MUST be specified as seed. The document justifies > > this by linking it to the lamps equivalent doc. But it is my understanding > > that there is a discussion there on allowing a format that encompasses seed > > or expanded key. I think it would be good to await the outcome there and > > ensure capabilities in the specifications are matching. > > > > Does that make sense, or did I misinterpret anything? > > > > Paul > > >
_______________________________________________ COSE mailing list -- [email protected] To unsubscribe send an email to [email protected]
