On Wed, Jul 22, 2020 at 4:46 PM Jim Popovitch via mailop <mailop@mailop.org>
wrote:

> On Wed, 2020-07-22 at 11:56 -0700, Marcel Becker via mailop wrote:
> >
> > On Wed, Jul 22, 2020 at 11:35 AM Jim Popovitch via mailop <
> mailop@mailop.org> wrote:
> > > On Wed, 2020-07-22 at 14:49 +0200, Sidsel Jensen via mailop wrote:
> > > > but if the effect is that it will drive up the adoption rate for
> DMARC then I am clapping my hands.
> > >
> > > "Once verified, the BIMI file tells the email service where to find the
> > > sender’s logo and the email service pulls that logo into the inbox."
> > >
> > >
> > > I don't think this is anything about DMARC, this is about inbox
> > > tracking.
> >
> > Um. No.
> > 1: DMARC is required for BIMI.
>
> Good, DMARC is good, but we don't need yet another standard to get DKIM
> and SPF into the wider use.
>
> > 2: A proper setup will proxy and cache the logo. eg: for us all you can
> track through BIMI is if our logo service is alive and well...
>
> I hope you understand that most providers don't care if your logo
> service is alive and well.  Surely we don't need a spec for that.
>

Once upon a time, there was the X-Face header... I guess we could have
updated that for some modern version
and sent it with every message.  For better or worse, a lot of mail clients
these days include an avatar
for the author, and acquiring that avatar can be complicated.  For the
larger mailbox services, there was usually
a proprietary way to set that avatar... iirc, for Gmail that meant creating
a G+ account for the email address
and setting the account profile photo.  Obviously, a standardized way to
specify avatars across clients is
better than proprietary methods... but that leads to a bunch of other
questions when it comes to
validation and preventing "abuse" of the avatar... whether or not BIMI is a
great answer to that is probably
still TBD.


> Whether you understand it or not, if a proxy or cache fetches your logo,
> you can get very valuable data about inbox hit rate data, eg tracking.
>

I think this highly depends on your MUA and it's model.  If Gmail fetches
and
caches the images, and there's no method in BIMI for "per user" images, then
tracking BIMI images is nothing more useful than "gmail received my
message."

If the fetching comes from the individual client running locally on some
personal
computer or mobile devices, the tracking level would be higher... at least
on the
"first" hit, assuming they did any caching at the client level.

This is all based on my assumptions about how this is all implemented, I
haven't read the spec yet.

Brandon
_______________________________________________
mailop mailing list
mailop@mailop.org
https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop

Reply via email to