Re: [Sks-devel] Proposal: Start verifying self-signatures

[Daniel Kahn Gillmor]

Thanks for the followup, Yaron--

On Tue 2015-05-19 18:18:44 -0400, Yaron Minsky wrote:
> Let's think about a simpler question: deletion.

I actually think after-the-fact deletion is a much harder question than
setting up filters based on some technical/discoverable aspect of
OpenPGP packets.  (though i suppose a new filter stated as something
like "everything except keys X, Y, and Z" could technically act as
deletion)

I think it would be a shame to twist ourselves in knots trying to come
up with a generic deletion policy (which seems like it would involve
quite a bit of political wrangling, centralized trust delegation, etc)
when adding a new filter (e.g. rejection of non-exportable keys [0]) is
much less fraught but no less important.

If we can solve the "how do we roll out a new filter?" question, that
puts us on the path to having effective machinery that we could then use
as a mechanism for deletion (if we decide as a community that's a good
idea).  but i'd rather not have to hash out the tougher question fully
(e.g. deletion policy) in order to roll out a new filter that doesn't
need those answers.

   --dkg

[0] https://bitbucket.org/skskeyserver/sks-keyserver/pull-request/20