[openpgp] Re: I-D Action: draft-ietf-openpgp-replacementkey-01.txt

[andrewg <[e-mail]>]

On 2024-11-03 13:14, Daniel Kahn Gillmor wrote:
> 
> With no chair hat on, my preference is for a mandatory imprint with an
> optional fingerprint (i think that's your option 3b).
...
> This allows an isolated implementation ecosystem that expects to have 
> no
> interaction with the keyserver network to not consume the extra space,
> but generally discourages it, which is the appropriate approach for the
> current scenario.

I am extremely skeptical of specifying this, because it violates one of 
OpenPGP's longstanding basic assumptions, that all keys have a single 
unique identifier. Even assuming that a particular cert will never be 
uploaded to a keyserver (and that's a big assumption), encouraging 
implementations to use context-specific identifiers, even internally, 
will almost certainly cause more problems than it solves. We're only 
talking about 32 bytes on the wire, and PQ signatures are going to be 
kilobytes in size, so it's proportionately not a big deal.

Stepping back, I'm concerned that a kludge that I threw together to 
cover a theoretical cryptographic shortfall has grown legs far beyond 
its initial intent. The imprint is only included in order to ensure that 
automated fallback from a v6 key to a v4 key has exactly the same 
cryptographic strength as going the other way. It is arguable that it is 
already overkill...

A