On 26/04/2022 14:48, andy yi wrote:
> On Wed, 13 Oct 2021, 14:38 s zhang, <
shezh...@gmail.com> wrote:
> Thank you for your reply.
> According to the documents and other
> information you gave, there are three
> digital signature algorithms shortlisted for
> the 3rd round of the nist pqc, namely
> CRYSTALS-DILITHIUM, FALCON, and Rainbow, but
> why does the chairman of abcmint claim that
> other than Rainbow, the other two algorithms
> are not feasible for digital currency and
> Internet applications.
>
This might be because the signature algorithms are not that able
to support threshold signatures, as currently defined. See...
https://eprint.iacr.org/2019/1060
Which kind of says Rainbow will be easier than the lattice based
schemes.
Personally I would not go with any oil-and-vinegar scheme now,
i.e. Rainbow, as there security claims seem to now be more debatable.
However, if one was prepared to tweak the lattice based signatures
[much like Schnorr can be seen as a tweaked form of DSA] then a
practical threshold lattice signature is possible
-
https://eprint.iacr.org/2020/1110
which is "Dilithium-like".
Perhaps when NIST re-open signatures they should have "ability
to thresholdize" as an explicit design criteria?
Nigel