On Tue, 16 Dec 2008 21:21:50 -0000, Hallam-Baker, Phillip <[email protected]> wrote:
> An update of the extensions draft that focuses exclusively on the X.509 > certs extensions is attached. > Note that even though this is intended to be an individual submission, > comments from others are welcome. The point of not making it a WG > charter item is that this work item is not sufficient by itself to > justify rechartering. > This has undergone substantial modification as a result of Paul's input > at the meeting. In particular I realized that there are significant use > cases for both 'call by reference' and 'call by value'. > Another major modification is that in this draft, ALL certificates are > encoded as a MIME application/pkix-pkipath object as described in the > TLS extensions RFC. So even if you have a single self-signed cert, you > have tio wrap it in a path. It is useul to bear in mind the possible use (at some possibly distant date) of DKIM for the purpose of signing Usenet Control messages (which are curently signed by a well-known but not well documented use of PGP signatures). The call by reference method would likely be used. BTW, why should a call by value occupy 4000 bytes of header? For sure, a well-attested PGP key can be stored in far less than that. And is there any inherent problem in using the DKIM signature mechanisms with PGP? A PGP key with a substantial web-of-trust behind it might be less hassle than organizing some accredited CA to sit behind the chain of a 509 certificate. Your mechanism is likely to be used for all sorts of special purpose applications, and it up to the commuminty involved in such applications to sort out what sort of keys they will use and how they are attested. > The reason for this is that in the context of DKIM, a self-signed cert > offers no real advantage over a DNS accredited key, other than an even > lower barrier to entry. As with SSL, I tend to think that self-signed > end-entity certs should be banned anyway, if someone wants to certify > themselves they should set up a CA and have one self signed issuing cert > and chain end-entity certs off it. I take your point about self-signed keys, which are roughly equivalent to PGP keys with an empty web-of-trust behind them. -- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Web: http://www.cs.man.ac.uk/~chl Email: [email protected] snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5 _______________________________________________ NOTE WELL: This list operates according to http://mipassoc.org/dkim/ietf-list-rules.html
