[Cryptography] Kyber PQC Key Exchange

[Adam P. Goucher]

> Yes, not every programmer needs to understand all algs.
> But personally I'm pretty convinced that no PQC algorithm
> is going to be very widely accepted unless it's possible
> for the top N% of programmers to understand that algorithm.
> (My current guess: N needs to be >=10, and we're nowhere
> near there now for PQC.)

I'd almost agree with this. The one exception is hash-based
digital signature schemes such as SPHINCS+:

http://sphincs.org/data/sphincs+-specification.pdf

The security assumptions are nice and simple; it just needs
a hash function that's pseudorandom and resistant to
preimage attacks.

Of course, SPHINCS+ is only a digital signature scheme, so it
can't be used for all of asymmetric cryptography: it doesn't
solve problems such as Diffie-Hellman key exchange or public-key
encryption.


Best wishes,


Adam P. Goucher