On 25 Dec 2016, at 1:14, Philip Homburg
<pch-ipv6-ietf-3(_at_)u-1(_dot_)phicoh(_dot_)com> wrote:
In your letter dated Sat, 24 Dec 2016 13:45:12 -0500 you wrote:
As I mentioned before, for me, the most usable option is avoiding
message modification of any kind, which leaves the origin DKIM
signature valid. I've not seen anyone comment on whether that's
workable for IETF WG lists (it works well enough for *this* list).
For me, the current behavior of the ietf lists is perfectly fine.
For me it’s not fine. And not just because I use gmail. If someone whose email
provider has a DMARC p=reject record sends a message to the list and half the
subscribers don’t see that message, the conversation gets disrupted even if the
other half sees it.
Check out the mailing list for Token Binding working group. The most frequent
posters are from Microsoft and Google, and I have to fish their messages out of
the spam folder every few days to keep up with the conversation. That disrupts
the entire working group, not just the people with the DMARC record or the
people whose provider respects DMARC headers.
It is very nice to see the mailing list in the subject. It is also
very nice to have a proper From header.
That's why I proposed a per-subscriber setting such that only those
who are actually stupid enough to reject or otherwise drop mail based
on DMARC have to see mangled mailing list traffic.
“Stupid”? Most of us have the choice of a corporate mail account where we have
no control over policy, and a public provider mail account where we also have
no control over policy. Yes, we can install our own mail server and set
whatever policy we would like. That is not a viable option for most people
There no technical reason for a one-size-fits-all solution. So let's
stop investigating those types of solutions.
A per-receiver setting is one extra step for new subscribers. That’s something
I’d rather avoid.
Yoav