Hello Hector, Brandon, Dave and John,

Thanks for the detailed responses to what may be naive questions that
pop up again and again... It is quite clear that SPF is more problematic
to make work through unknown relays and lists than DKIM, and so that my
proposed work on Lenient DKIM is a better thing to work on.

As for responding to such questions, it is indeed helpful to know *why*
things were decided against, especially if this was 10 years ago. 
History should not repeat itself, but the grounds for a decision are
useful to weigh again every now and then.  This is one of the reasons
why I was asking around.

It's interesting to hear that rolling back of Received: headers is done
at Google [apparently with /128 prefixes for IPv6, and resulting trouble
with private address changes] and the reason I find it interesting is
that we're designing an identity system that would also manage external
aliases, so it could do the thing of which we discussed that [at least
some] users might use to shoot themselves in the foot.

Very clear in all this is that SPF is riddled with problems and should
not be included lightly.  SRS puts a strain on intermediates and won't
solve things IMHO.  I wish I could say otherwise for DKIM, but that too
breaks more than I like, and so I am really really looking forward to
the next ARC draft and hope to also come up with a MIME-aware
canonicalisation proposal in proze/draft and poetry/code.


Thanks,
 -Rick

_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to