From: "Olson, Margaret" <molson(_at_)roving(_dot_)com>
...
Combining context/content labeling with permission labeling might get
something more useful, for example:
Person to person
Transactional - implicit permission. If I buy something, it's fair for the
vendor to assume I want the email verification.
...
I probably left out several useful combinations. This is really just
permission labeling with a little bit of context, and some combinations are
clearly nonsense (transactional - no permission, for example).
Those categories or any others you prefer are fine, as categories.
This gives the MUA (or MTA) a short list of filters that an end user might
reasonably be willing to deal with.
Please say how you or anyone else would actually use those categories
for incoming mail. Of course bulk mailers including organizations like
Roving Software as well as legitimate web merchants would be glad to
use labels to make their messages more likely to be delivered and read,
but those are uses of labels by senders instead of by targets. In the
real world practically no mail target is going to bother to configure
filters to pick and choose types of messages from particular senders.
Everyone wants some messages of almost any category from some sender.
What none of us want is any and all mail from some senders, and any
and all unsolicited bulk mail from most senders.
The only two categories that matter in practice are "spam" and "not
spam." That none of the content label proposals have an "unsolicited
bulk email" category is emblematic of the irrelevance of content labelling.
We're probably stuck with the transactional messages that are also full of
ads - I don't know about you guys but almost every bill I get is stuffed
with ads of some sort.
That's part of why in the real world the labels won't matter to mail
targets. Senders that cannot be trusted to send only what targets want
will "fib" on their labels. "Transactional" will be used where "ad"
would more accurate. Some advertisers will send "person to person"
notes to 30,000,000 of their closest friends. They already do that
with Subject lines. You could write laws against that sort of thing,
but then you'll have advertisers insisting that since their messages
are addressed to individual people and are written by real people whose
names are on the messages, they really are "person to person."
However, that's not the main problem. Users do not like setting up
filters. They simply will not bother to configure their software to
accept "transactional" messages but not "ads" from Roving Software.
They will block all or none of your messages. It doesn't matter
whether you use per-message content labelling or you do as Topica has
tried, labelling by domain name or Habeas mark. Unless essentially
all of your messages are wanted, all of your messages will be blocked
based only on the fact that you are the sender (like the most recent
edition of the weekly drivel that your organization tries to send me).
Spam won't stop being spam and won't stop being filtered by having an
additional label.
Vernon Schryver vjs(_at_)rhyolite(_dot_)com
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg