Re: [openpgp] considering other OpenPGP work

[Andrew Gallagher <[e-mail]>]

On 4 Apr 2024, at 22:08, Daniel Kahn Gillmor <[e-mail]> wrote:
> 
> In the same list of documents, i see this individual draft about
> replacement keys, from Daphne Shaw and Andrew Gallagher, which appears
> to cover the idea of "superseded keys" mentioned in the milestones:
> 
>   https://datatracker.ietf.org/doc/draft-gallagher-openpgp-replacementkey/
> 
> Daniel, Daphne, Andrew, are these documents anything that you would be
> interested in considering for working group adoption?  If so, please
> take a minute and write a paragraph or two to the list here about the
> draft and why you think it would be useful for us to consider it as a
> group.

Yes, I was planning to bring this up at the last meeting but disease and lack of time prevented it. :-)

There is currently no standardised method for a key owner to refer their correspondents to a new key that the owner would prefer to use. This will become particularly important as we manage the v4-v6 transition. Daphne wrote a draft several years ago that I believe definess an elegant solution to this problem, and we have worked together to bring it up to date with modern practice.

The draft is very lightweight - it defines a signature subpacket with a similar format to the Issuer Fingerprint and Intended Recipient subpackets, but it is also capable of holding an explicit null value. It is intended for use on a self-signature only, to indicate the fingerprint of a key that the key owner wishes their correspondents to use instead. It is specifically not restricted to revocations, and so allows the new key to be specified in advance of the old key’s expiry date. The explicit null value is used to indicate that there is no replacement key (as opposed to being merely undefined). No trust value should be inferred from the mere existence of a Replacement Key subpacket.

We believe this is a key component in the migration process to v6 and/or PQC, as it allows clients to automatically upgrade their correspondence to use more modern keys without user intervention (contingent on the usual trust calculations). We therefore ask the WG to adopt it for publication on the standards track.

> If anyone else has other work that they would like the working group to
> consider as a possible work item in the near future, now might also be a
> good time to propose it, so that the working group can consider it.

Aside from the above, I have been working on updating Daphne’s HKP draft. This is at a less advanced stage than the above, but is more complex and would benefit from additional eyes.

https://datatracker.ietf.org/doc/html/draft-gallagher-openpgp-hkp

I would like it to be adopted once the WG feels it has capacity. In the meantime, if people have time to read the latest draft and would like to speak up on the issue tracker it would be most welcome, as there are several outstanding questions.

https://gitlab.com/andrewgdotcom/draft-gallagher-openpgp-hkp

Further down the line, I would personally like to see progress on revocations and 1pa3pc. :-)

A