I agree with your second point without conceding the first.  We were 
experimenting with different possibilities in a collaborative development 
stage.  This wasn't supposed to be a fixed final spec, ready to be "implemented 
correctly".  The development stage is different from coming along and 
implementing a static spec later.

And I don't think this line of discussion is leading anywhere productive.  I 
continue to hold that the characterisation of that hackathon in your draft is 
unfair.

Bron.

On Mon, May 11, 2026, at 19:39, Vittorio wrote:
> Hello Bron, a specification simple enough to be implemented correctly 
> during a hackathon without "changing the spec multiple times as we 
> hacked" would not have produced those failures. Complexity is not a 
> neutral property, it correlates directly with implementation error 
> rates, regardless of the specific errors encountered.
> 
> Regards
> Vittorio Moccia
> 
> Il 2026-05-11 02:32 Bron Gondwana ha scritto:
> > Again, I continue to object to the characterisation of the
> > DKIM2-interoperability work done at the hackathon of IETF124.  I was
> > there, you were not.
> > 
> > The issues we ran into were partly due to changing the spec multiple
> > times as we hacked, and partly due to a bug regarding whether line
> > endings were included.  Neither of those issues would be mitigated by
> > the changes you propose, and neither would be an issue when have a
> > stable specification with good test vectors.
> > 
> > Bron.
> > 
> > On Sat, May 9, 2026, at 17:07, Vittorio wrote:
> > 
> >> All,
> >> a revised version of the deployment profile has been submitted.
> >> 
> > https://datatracker.ietf.org/doc/draft-moccia-dkim2-deployment-profile/
> 
> _______________________________________________
> Ietf-dkim mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
> 

--
  Bron Gondwana, CEO, Fastmail Pty Ltd / Fastmail US LLC
  [email protected]

_______________________________________________
Ietf-dkim mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to