Re: [TLS] On counting

[Adam Caudill <[e-mail]>]

On Jun 29, 2014, at 6:49 PM, Paul Hoffman <[e-mail]> wrote:

> If you are saying that only people who are speaking on the list want to have Curve25519 be adopted and everyone else doesn't and thus the first group is a small minority, that's a gross assumption, not a fact. You have been active in the IETF long enough to know that many people stay silent not because they agree with the way things are, but because they are sick of the tone of the discussion.

There are certainly more people that want to see curve25519 adopted than those that are speaking up; myself included. I wouldn’t advocate requiring it, but there are cases where it would be useful. If it makes sense for an implementation, implement it, if not, don’t - but there’s no reason to prevent it from being used in the cases where it would add value. It performs well, it’s generally accepted to be secure - otherwise I’m sure the reaction from the CFRG would have been quite different.

This isn’t just about being non-NIST, it has real value from a performance and implementation perspective. Personally, I see no reason not to adopt it - though I wouldn’t go so far as to require it be implemented (though I think it should be in many cases), or state that it should be a preferred option.

-- 
Adam Caudill