On Sat, Mar 28, 2015 at 11:56 AM, Daniel Kahn Gillmor
On Thu 2015-03-26 15:58:57 -0500, Phillip Hallam-Baker wrote:
Yep, I don't actually rate a keysignature as being worth anything
until it is enrolled in a TRANS like log.
I think this perspective is dangerous to privacy, if we want to be able
to support non-public certifications.
I often certify people's keys publicly (or i make OpenPGP certifications
and send them to the subject and let them decide whether to publish them
or not). But i also think it's important to be able to make a
non-logged, non-public "letter of introduction", to be handed off when
needed. OpenPGPv4 already supports this in the form of non-exportable
signatures. The UI/UX for this is abysmal in most clients today (anyone
with UI/UX cycles to spare who wants some brainstorming ideas about how
to improve this, please talk to me), but it would be a real shame to
change the protocol in such a way to rule this technique out completely.
Parties who are globally relied-upon (e.g. the X.509 CAs that everyone
implicitly "trusts") should definitely be publicly logged.
But not everyone who certifies is in (or should be in) that position;
some of these relationships are private, and we should not force people
to publish them.
By that I mean fixed in time. I agree that it does not need to be
public. Only the hash needs to be enrolled.
openpgp mailing list