Re: [TLS] Review of draft-ietf-tls-openpgp-keys-08

On 5/16/06, Pasi.Eronen@nokia.com <Pasi.Eronen@nokia.com> wrote:

> > > Somewhat substantive comments:
> > > 1) Section 3.3: The TLS Certificate message defined in RFC4346
> > >    contains a list of X.509 certificates, while this section allows
> > >    only a single PGPKey structure. IMHO we should keep this aligned
> > This was a deliberate change. The rationale behind this change was
> > that there is no point in sending more than one pgp keys. There is no
> > notion of certificate chain in openpgp, thus having a list of
> > certificates would confuse on what should be sent.
> If Alice signs (certifies) Bob's key, and Bob signs Carol's key,
> isn't that a certificate chain? (The whole notion of "web of trust"
> seems to imply a certificate graph, which again implies chains...)

Indeed that's true.But it is also not very practical to include
signers in the key sent.
That is because a key might be signed by tens or hundrends of keys, that are not
compact either. I also see not much usage of the extra keys. The
typical way for the receipient to verify an openpgp key is by
retrieving by himself the required chain, if needed.

A similar matter was discussed in the early days of the draft... i
just found a link:
http://www.mhonarc.org/archive/html/ietf-openpgp/1997-12/msg00105.html

regards,
Nikos

_______________________________________________
TLS mailing list
TLS@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/tls