Re: [openpgp] The Argon2 proposal seems incomplete (Draft 6)

[Daniel Huigens <[e-mail]>]

On Thursday, August 4th, 2022 at 11:43, Bruce Walzer wrote:

> You could do this:
> 
> > The values of t, p and m, from section 4 of [RFC9106] are RECOMMENDED.

Yeah, that might be better. Do you happen to feel like making a MR?

> ... but section 4 makes several suggestions for those values so this
> wouldn't be an actual recommendation. It implies that it might be best
> to tune the parameters which would be obviously absurd for the OpenPGP
> case.

Why would it be obviously absurd? What if an application indeed has one
of the use cases laid out at the beginning of section 4, so that one of
those would be appropriate? Or, what if it has a use case for symmetric
encryption on a single device, and wants to select values for that
specific device, as laid out at the end of section 4?

The current vagueness in the text allows implementations some wiggle
room, and given that OpenPGP is meant to be a general specification,
that might be a good thing.

Best,
Daniel