On October 8, 2021 4:31:03 PM UTC, Dave Crocker <[email protected]> wrote: >On 10/8/2021 9:28 AM, Scott Kitterman wrote: >> Agreed. I was confused because it appeared to me that you were directing me >> there for an answer about DKIM signing and I couldn't find it. > > From your note I thought you didn't know about the spec. (No, that >doesn't seem like a reasonable believe on my part, but I'm still working >on my second cup of coffee.) > > >> In shorthand terms, Author is the opposite of Sender. In the existing >> Sender paradigm From is constant, the mediator would add Sender, which would >> result in sent by Sender on behalf of From. Under this proposal the >> originator includes both From and Author, the mediator mangles From and the >> result could be sent by From on behalf of Author. Is that right? > >One of the reasons I pointed to the draft is that it discusses the >history and semantics of Sender vs. From and makes the case the DMARC >forces From to be Sender, but not really From any more. Author seeks to >recover a purely From semantic.
Thanks. That matches my understanding. My vague recollection is that the reason not to use Sender (implicit or explicit) as the key for ADSP and later DMARC was concern that some MUAs didn't display the explicit Sender (mostly Outlook Express, IIRC). The original Yahoo! DomainKeys had some sort of a policy component that keyed off Sender. I haven't gone back and looked anything up to be sure, so no promises. Maybe that was the right answer all along. Are MUAs that don't display Sender still a concern? Do we care? Maybe keying off Sender instead of From gets us to a similar place without requiring upgrades to every MUA in existence? Scott K _______________________________________________ dmarc mailing list [email protected] https://www.ietf.org/mailman/listinfo/dmarc
