Re: [dispatch] Updating DKIM for stronger crypto

[Martin Thomson <[e-mail]>]

On 21 Mar. 2017 15:01, "John R Levine" <[e-mail]> wrote:

In looking at this, we should probably dispatch this cross-area to
> CURDLE.
>

Please, no.  There's nothing interesting about the choice of algorithms,
only about how to update the DKIM spec to say what to do.


That doesn't seem like a good reason. I just spent a few minutes with RFC
6376 and it seems obvious that the changes required are minimal as long as
you can conform to the well-established signing API that everyone -
including DKIM - uses. You write a new draft that defines a new codepoint
and the algorithms that the codepoint represents and that is it.

That describes exactly what CURDLE does.  Frankly, the process is pretty
mechanical, if you want to pad a resume with RFCs, it is a good way to get
started.

If there is some way in which DKIM really *is* special despite outward
appearances, I'd be very interested in hearing about it. Requiring a small
public key representation isn't especially interesting or challenging.