At 06:51 PM 3/4/2003 -0500, Paul Judge wrote:
spam can not be expressed as a boolean
While I suspect the statement might be true, I would counsel you to avoid
absolute statements this early in the game. They shut down exploration in
the deprecated paths, research that might be of value.
If we can describe classes of traffic, such as perhaps providing keywords like
PORN
HUMOR
OFFER-TO-SELL
PERSONAL-MESSAGE
TOPICAL-DISCUSSION
OFFER-TO-ILLEGALLY-LAUNDER-EXCESS-NIGERIAN-CAPITAL-FOR-FUN-AND-PROFIT
TROLL-FOR-INTEREST-IN-OFF-TOPIC-DISCUSSION
(well, maybe not the last two)
one could imagine people being able to identify their traffic correctly and
being motivated to do so. While "you sent me unsolicited commercial email"
is not an accusation (oops, statement) that can be objectively evaluated
because the spammer might have a different definition of "solicit" than you
do (those who spam me sure have a different definition than I do; they tell
me I actually solicit this crap), the statement "this is a solicitation to
view a web site containing material of an adult nature" is objective enough
that folks who send them seem to be able to identify them - and mention the
adult nature, the fact that they are 18, or various products such as
NetNanny in the course of their solicitation.
Anyone who sees a need to include "to unsubscribe ___" or "I claim that you
opted in because some mailer at the other end of the earth got loaded up
with your email address" might find it in their interest to correctly
describe the load they are offering. A mail relay server might, on that
basis, only choose to handle traffic that claims to be of type
TOPICAL-DISCUSSION, and might state in its mailing list description that
the filter was being applied, or might include a note to that effect in its
bounce message.
You still have the question of whether folks sending traffic of type PORN
might decide that their mail is always TOPICAL-DISCUSSION, but I suspect
that we will always have the problem of non-cooperative senders of mail.
What I have in mind, of course, is something similar to what Dan Bernstein
does - if you send him an email, you will get a very obnoxious note that
says "if you really want to dent my consciousness, you need to send your
email again and tell me that you will pay me $250 if I read it and decide
it is spam." But what I would propose is some form of system in which you
describe the contents of the email in terms that a boolean filter (a filter
that applies a boolean expression and does something if it is true and
something else if it is false) can evaluate. What is needed is a vocabulary
of message types.
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg