On October 8, 2021 5:43:43 PM UTC, Dave Crocker <dcroc...@gmail.com> wrote: >On 10/8/2021 10:34 AM, Scott Kitterman wrote: >> I think it's fair to consider that Sender is at least implicitly always >> present. > >In the abstract, according to the spec, yes it is. In practice, >arguably it often is not. > >To the extent that Sender should indicate the 'operator' of the >mechanics, distinct from the 'author' of the content, the Sender >information often is not present. > > From a DMARC perspective, what one would have wished for is for all >mail to always have the Sender field included explicitly, and DMARC work >off of that, independent of the From field. > > >> Having a MLM add Sender and not munge From is a far better UX than the >> munged From. Lots of software already supports it too. > >That requires trying to change DMARC. That's not going to happen. > > >> Would it make sense, perhaps, to key DMARCbis off Sender (i.e. Sender if >> present or From if no explicit Sender)? If that makes overall sense, it >> would substantially simplify the MLM's problem. > >Do the transition matrix for that, showing interactions between changed >and unchanged DMARC participants. (I did that during this round of >effort that lead to the Author proposal.) The result isn't viable.
Thanks. My conclusion then is that I don't see anything particularly problematic with Author beyond every MUA in the world needs to be updated, but given the undeniable reality of From rewriting I can see that it might eventually be helpful. I'm still on the notion that there isn't a single viable solution to the DMARC "mailing list problem" and we should focus on documenting things senders, mediators, and receivers can do to lessen the severity and frequency of it coming up. Scott K _______________________________________________ dmarc mailing list dmarc@ietf.org https://www.ietf.org/mailman/listinfo/dmarc
