At 12:58 PM 4/4/2003 -0500, you wrote:
There is a simple reason for trying to 'define spam' in the context of this
group. It is not related to anything to be construed as a 'globally'
acceptable definition per se. It is merely intended to be a tool for,
ostensibly, our use in evaluation of the proposals. IMHO, we should focus on
such a solution for a definition. It will not be worthwhile, I think, to
continue in debate on a less precise definition. In some cases the
definition
describes the 'scope' of a problem and uses that to develop a solution
(reduction of impact, scope, etc.), in others a precise term has been used to
identify a problem and formulate a solution to eradicate the defined object.
In this case we should try to define 'it' in one of those two ways so that
the
effect of a proposed solution can be measured objectively by the group for
research purposes, and tested in interoperable implementations developed
(nominally) outside of the group (in the IETF) [RFC2014, BCP8].
Yes, but that's why I favor sub-definitions. The Venn diagram would show
sub-definitions mostly covering the entire class of "spam", with some
sub-definitions disjoint, others overlapping. If there's a sub-definition
that associates well with a specific method of fighting spam then I'd think
the evaluation of that method would be based on its effectiveness for that
sub-class of spam, not on how well it worked on spam as a whole. I'm
assuming the possibility that the ASRG solution won't be unitary, that it
could be a combination of approaches. If a unitary solution is found the
discovery will be (perhaps) made by observing that the associated
definition covers all spam and that the method is effective against the
entire sub-class of spam associated with the method. Rather than first
setting the difficult goal of finding a master spam definition and then
finding a method that works against all that spam the approach would be to
identify subclasses of spam that have anti-spam measures associated
with them.
Split the problem into parts, deal with each part, solve the problem in pieces.
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg