[ietf-dkim] treewalking vs nxdomain

[Steve Atkins]

On May 1, 2008, at 7:41 AM, Al Iverson wrote:

> Is anybody else of the opinion that NXDOMAIN recommendations (at
> least) would help fill a perceived gap if treewalking is left out? I
> do feel that way. I think the point is to help provide receiver
> guidance on what to do with failed mail, and providing policy guidance
> for non-resolvable or unconsidered FQDNs or domains seems to be the
> goal of both bits of functionality.
>
> It feels to me as though an NXDOMAIN-related recommendation is less of
> a hack and far more suitably implemented than treewalking, and I think
> it might end up being a good alternate path.
>
> Sorry to want to boil the oceans, Dave.

I'd agree. More, actually, I think that the algorithm is entirely  
broken if
there's not something similar to the NXDOMAIN check in there, while
the treewalking does not map well onto how DNS is used in the real
world, shifts complexity to the wrong parties and increases the network
overhead of every email message received.

Cheers,
   Steve