signatures by a key that has since expired

[David Shaw]

--ReaqsoxgOBHFXBhH
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Jul 18, 2003 at 11:41:50PM +0100, Neil Williams wrote:

> 1. How should I handle signatures on my key where the signing key
> has since expired? (The key was valid at the time that the keyholder
> signed my key.)
>=20
> My key is on most keyservers, so I tried revsig but was only offered
> self-signatures. (I guessed as much but thought I'd try to make sure
> before someone suggested it!)

That's correct.  You can only revoke signatures you made yourself
(whether on your own or someone elses key), hence the self-sigs.

> 2. Is it expected that those people who use keys with limited
> expiries should take the responsibility for signature maintenance on
> other keys?

This is really a cultural issue.  The answer is "maybe" :) Seriously,
it depends on the signer.  If the signer chooses to update their
signaures, that is up to them.  There is no requirement in the
protocol either way.

> If that isn't possible or if it isn't wise to allow me to use revsig
> on someone else's signature of my own key, should there be / is
> there some way of filtering out signatures made by keys that have
> expired?

The idea of a "key pruning" has been suggested a few times, most
recently by Michael Young.  The idea is that GnuPG would strip off any
expired/bad/untrusted/etc signatures, leaving the key with only valid
and current signatures.  It's a good idea, though the advantage is not
universal.  A keyserver refresh would naturally replace all of the
signatures, valid or no, so the pruned key would be most useful for
people who wanted to store their key on a web site somewhere and
wanted to keep the file size small.  I believe Michael had a different
need though.

> 3. Could --check-sigs illustrate this in a similar was to revoked
> signatures?

Good idea.

> 4. What is the general consensus on the validity/trustworthiness of
> expired key signatures? (after all, the signature itself is still
> valid.)  What is the consensus on validity etc. of the expired key
> itself?  Why are expiry dates used and what is supposed to happen
> when the key does expire?

My feeling is that the signature is indeed still valid, but it has no
meaning any longer.  Look at it this way: the signature is a statement
made by key A about key B.  The statement may well be valid, but the
expired key that made it is no longer valid - a valid statement by an
invalid key.

This can be an odd situation, since it means that very popular keys
that are strong points of the web of trust must never be allowed to
expire, even if the key owner no longer uses that key.  It could be
argued that this is either a good or a bad thing.

> 5. Are signatures made by a key that has since expired removed from
> the web of trust calculations?

Yes, just like signatures made by revoked keys.

David

--ReaqsoxgOBHFXBhH
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3rc1 (GNU/Linux)
Comment: Key available at http://www.jabberwocky.com/david/keys.asc

iD8DBQE/GXLN4mZch0nhy8kRApw7AJ4whAZ87jwzYSlbMzTrNUpneKJVEACcDbpu
Cf/jYlxSWwpFjuEv+8dlJc8=
=Yrb9
-----END PGP SIGNATURE-----

--ReaqsoxgOBHFXBhH--