[ietf-dkim] 1193 considered harmful

[Dave Crocker]

Barry Leiba wrote:
>> How does this address my concern? This looks like my current receiver
>> would fail with the new signature format. That's not backward compatable.
> 
> All verifiers already have to change, to support SHA-256.

Sounds like this is another exchange talking past each others' meanings.  At the
least, I am increasingly confused by which concern people are responding to.

So:

1) Yes, an IETF DKIM verifier must be different from a pre-IETF verifier.
However we have so far preserved the ability of a pre-IETF signer to work with a
post-IETF DKIM verifier. (So, Barry's statement is true, but I'm not sure it
addressed the concern.  Or am I, yet again, missing the point?_


2) When we talk about compatibility, could we start using some sort of notation,
to make very clear what type and direction we have in mind.  For example, I
believe the over-the-wire concierns divide beteen::

   a) pre-IETF DKIM -> post-IETF DKIM compatibility (pre2post)

   b) post-IETF DKIM -> pre-IETF DKIM compatibility (post2pre)

So far, we have preserved pre2post compatibility but do not have post2pre.

d/
-- 

Dave Crocker
Brandenburg InternetWorking
<http://bbiw.net>