Re: Thoughts on an OpenPGP header?
On 2004-12-06 22:26:11 +0100, Simon Josefsson wrote:
> if the recipient sent the incoming message, and it was signed or
> encrypted, use that key.
What would the *sender*'s key used for *encryption* be? We're
dealing with public-key cryptography, so encrypting a message does
*not* naturally involve any key material of the sender.
> search for a key with a matching uid.
That should still be the first step, as it's the one where we can
actually check whether some key is valid, i.e., whether we know one
way or the other that it belongs to the intended recipient.
> Ask the user which key to use, using the one mentioned in the
> OpenPGP header as the default choice.
Once again, an OpenPGP header won't help to establish whether or not
a certain key validly belongs to an e-mail address. The only thing
you can use it for in this context is to prioritize among keys that
are known to be valid, and to identify keys for download whose
validity is then verified out of band or through the web of trust.
(This validation step will also take care of the 0xdeadbeef
attacks.)
I'm a little concerned, by the proposal to include further
information about the key in this header: All the information
mentioned is duplicated inside the key material, and there in a
secure form.
Probably, the most useful approach for this header is to specify its
semantics as a simple list of URIs for OpenPGP key material.
Using the header as an indication for PGP/MIME abilities might also
be a good idea.
--
Thomas Roessler · Personal soap box at <http://log.does-not-exist.org/>.