On Monday, March 31, 2003 3:49 AM, Jon Kyme
[SMTP:jrk(_at_)merseymail(_dot_)com] wrote:
Besides that, I think in your last sentence you are describing
a policy enforcement mechanism based upon a precise (or im-precise,
or subjective) definition of a 'class' of messages.
Yes, kind of. A mechanism has to rely on the classification tools available
at the classifying entity. This allows for a universal definition
in terms of local classifications, and avoids the (probably impossible)
question of how to create a single universal definition of a "spam
message".
Given a few terms universally accepted ("acceptable", "unacceptable",
"neither") and classification "engines" which need not be well-known,
transparent or
reliable.
You can get classifications like:
"messages classified as unacceptable by
billy-bigarms-patent-spam-detector".
"messages with senders on the recipients blacklist"
"messages originating-from-smtp-client-ip(10.0.2.3)"
etc...
Obviously, enforcement can only be performed by entities with access to the
engines specified. This is a "good thing".
I agree. So is there consensus here on the idea that enforcement make take
many forms that are based on the principle of an identifiable construction
(class) against which the policy may be applied/checked?
Classification of a message is not impossible.
(There are many deployed systems that do this with varying degrees of
"accuracy") Expression of a subjects preferences is not beyond us.
We don't need to define spam. All we will ever need to do is say that
"this set of message classes does | doesn't have consent for delivery
to
this subject."
--
--
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg