Re: [openpgp] Call for adoption of draft-gallagher-openpgp-replacementkey

[Simon Josefsson <[e-mail]>]

Andrew Gallagher <andrewg=[e-mail]> writes:

> On 7 Apr 2024, at 18:34, Simon Josefsson <simon=[e-mail]> wrote:
>> 
>> Is this intended to replace or augment human-written key transition statements?
>
> Yes.
>
>> Doing so would be useful, and
>> quite important to ease key roll-over to PQC keys, but transition
>> statements are usually not published when keys are expired or revoked.
>
> Indeed, this subpacket gives a key owner the means to make such a
> transition statement in machine-readable form, and distribute it
> through the customary key update mechanisms.

That seems useful, and I think it would be useful to address that
problem, however key transition is a complex problem as it interacts
with trust settings which PGP historically largely have left outside of
scope for the protocol.  This reinforce my feeling that there is a gap
between the current state of the draft and whatever the goals may be,
and that the effort would be helped by taking a step back and describing
the goals more clearly first.  I do think it is wortwhile to explore,
and will try to help by participating in discussions.

One aspect is that a machine readable form is only useful if machines
can make useful decisions based on the data.  Is that the case?  There
are security concerns with automated key transition mechanisms.
Consider if someone manages to get me to sign a replacementkey binding
to a new vulnerable RSA1024 key, how to revoke that statement?  How to
evaluate it on the recieving side?

/Simon