Re: [openpgp] German BSI, PQC for OpenPGP in Thunderbird,

[Daniel Kahn Gillmor <[e-mail]>]

On Mon 2021-06-28 12:36:17 +0200, Justus Winter wrote:
> I think we should revisit the way we store cryptographic artifacts in
> OpenPGP.  Unfortunately, neither SOS nor SBS address the issue of
> potentially large PQ artifacts.

As the person who coined "SBS", I want to note that it is *not* intended
to be a design that will work for future algorithms.  It's merely a
handle to describe the particular idiosyncratic structure used in some
cases already for 25519.

While SBS is "MPI-like", we already do have non-MPI-like data structures
stuffed into the "algorithm-specific" parts of some key material: in
particular, the ECC OID is encoded with a single octet length, which
means it cannot be parsed correctly by any implementation that is
ECC-ignorant but blindly expects to extract a series of MPI-like
objects from a public key packet.

For future algorithms (including anything PQ) i'd hope that the future
spec would explicitly declare a distinct data structure for the
algorithm as necessary.  I don't think we need a generic data structure
format given the range of possible requirements.

       --dkg