On Sat, 2007-01-20 at 03:07 +0000, John Levine wrote:
> >>6. Verifier Actions
> >>
> >>  Since a signer MAY remove or revoke a public key at any time, it is
> >>  recommended that verification occur in a timely manner. In many
> >>  configurations, the most timely place is during acceptance by the
> >>  border MTA or shortly thereafter. [In particular, deferring
> >>  verification until the message is accessed by the end user is
> >>  discouraged.]
> >>
> >>This precaution should be removed!!
> >
> >I disagree with Doug and agree with the wording in the current document.
> 
> I'm with Paul, I do not want to reopen the arguments about how long a
> verification key should or shouldn't be around.

Why strengthen a bad statement that attempts to declare DKIM is to be
done only at the MTA?  How can DKIM's protection be extended to the MUA?
The concept that policy will block look-alike attacks at the MTA is
highly flawed, and even more so once EAI becomes commonly used.  DKIM is
_designed_ to be invisible.  DKIM's protection requires annotation be
added.  The most secure place to add this annotation is at the _MUA_.

How can an MTA know which email-addresses a recipient trusts?  Checking
the validity of every signature will leak information in the same manner
pulling in separate graphic images.  A bad idea.  Why should the DKIM WG
insist upon making this mistake?

Marking messages at the MTA is afforded none of the protections provided
by the DKIM signature.  Why open such a security hole?  This added
statement goes to the heart of the questionable thinking that surrounds
sender policy protections.  This is a conversation that should be
carefully reviewed and well understood.

-Doug   

_______________________________________________
NOTE WELL: This list operates according to 
http://mipassoc.org/dkim/ietf-list-rules.html

Reply via email to