On Tue, Oct 14, 2025 at 01:44:22PM -0700, Eric Rescorla wrote:
> On Tue, Oct 14, 2025 at 1:30 PM Nico Williams <[email protected]> wrote:
> > I should add that if the necessary codepoint registrations are
> > Specification Required (as they are here),
> 
> Indeed the code points have already been assigned for MLKEM.

Right.

> > then a policy of non-
> > publication of PQ-only suites would have no real effect unless the
> > registration policy is changed to be IESG Protocol Action.
> 
> This change seems like it would be very unlikely, given that we
> changed to the current policy precisely to address this this
> kind of debate (which has obviously not succeeded completely).

Nonetheless the industry has been burned by Dual_Ec, as you know, and
this looks like it could possibly be repeat.  Given that we can't really
forbid pure-PQ, it should at least come with warnings to use it mainly
where performance is essential and mainly only in, e.g., corporate
networks.

Thus perhaps we should indeed publish pure-PQ as Experimental and with
caveats.  Granted, people see "RFC" and think "standard", and don't look
inside, but since it will ship, we might as well have an Experimental
RFC.

[0] And the extended random that enabled Dual_Ec's use as a backdoor.

Nico
-- 

_______________________________________________
TLS mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to