[dmarc-discuss] A bit quiet?
aiverson at spamresource.com
Mon Oct 26 09:08:34 PDT 2015
On Mon, Oct 26, 2015 at 12:45 AM, Roland Turner via dmarc-discuss
<dmarc-discuss at dmarc.org> wrote:
> Al Iverson wrote:
>> From my own perspective, I'm unclear on how well this will work. I
>> assume the chain process is based on addressing anything thrown at at
>> it; mailing list posts going through mail forwarding; ARC on both
>> would in theory keep authentication intact and prevent p=reject policy
>> rejections. But we're talking the 1% of the 1% (of the 1%?), it feels
>> like the use cases might get more and more far out.
> I'd suggest that what ARC solves - if it works - is the entirety of the problems for forwarders who are willing to cooperate but nonetheless wish to modify messages sufficiently to break DKIM, which remains the largest class of inadequately solved problems with DMARC. (Note that the current low fraction of p=reject mail is not hugely important; as the DMARC breakage cases disappear, a growing fraction of email can and will be subject to p=reject.)
> There remains one unsolved significant case, that of independent origination ("share this link") which, I suspect, will be permanently beyond reach for interoperable protocol standardisation (it depends entirely upon trust by receivers and not at all upon protocol mechanisms).
You get a gold star for thinking of a use case I had not considered!
"Share this link" AKA forward to a friend/FTAF/F2F. Yeah, potentially
an issue, but was already kind of an attractive nuisance when a
marketer tries to incentivize it (and thus really, really wants to
track it). If you don't care so much about tracking conversions/sales
that come after, you could probably just replace links to
"mailto:friend at friend.example?subject=widget&body=check_out_this_widget."
Then the mail starts from the initiator's MUA instead of from third
Al Iverson - Minneapolis - (312) 275-0130
Simple DNS Tools since 2008: xnnd.com
www.spamresource.com & aliverson.com
More information about the dmarc-discuss