ietf-dkim
[Top] [All Lists]

Re: [ietf-dkim] assertions, was on DKIM as an anti-spam measure

2005-08-17 20:36:33
1. they increase the liklihood that their messages would reach the
intended audience because recipients have explicitly allowed it to
bypass their spam filters.

(because those recipients specifically said they want particular kinds of
advertising.  there are people who need to receive narrowly-focused
advertising or find it useful)

People I know who want narrowly-focused advertising go out and sign up
for it.  This does not need assertions, DKIM, or anything else.  DKIM
and reputation systems could be useful for people to report whether
mailers actually do what they said they would, but that doesn't need
assertions either.

In a system like usenet where everyone sees every message, assertions
could be useful to pluck relevant messages out of the stream as it
goes by, although topic tags aka newsgroup names seem to have done the
trick for the past 25 years.  In e-mail, this would make sense in a
model where mailers spam out their stuff to every address in the known
universe, recipients look at the flood as it comes in, reject the
99.99999% that they don't want, and accept the trickle with the right
tags.  I sure hope that's not what you have in mind.  So once again,
What Problem Are You Trying To Solve?


2. there were a penalty of some kind for not making assertions in some
circumstances (e.g. US laws requiring some kinds of messages to be
labeled as spam or sexually-explicit)

Gee, by an amazing coincidence, someone with exactly the same name as
you has published notes about what a bad idea porn tags are.  When
Korea mandated spam tagging, it opened the spam floodgates and ruined
e-mail in Korea, perhaps permanently, even though the spam tag law has
been repealed.  If you don't believe in standardizing unsound existing
practice, why in the world would you want to encourage this?

R's,
John

_______________________________________________
ietf-dkim mailing list
http://dkim.org

<Prev in Thread] Current Thread [Next in Thread>