On Dec 30 2004, Peter Kay wrote:
1. The vendor claims are false. In the real world, you still get lots of
undesireable email.
This is true. Vendor claims generally mean nothing and tend to be
misleading to sell their wares. There are many issues, such as:
1) Nobody verifies a vendor's claims except anecdotally, because it's too
much work or needs large resources. So you get junk like this instead:
http://www.zdnet.com.au/insight/software/0,39023769,39172027,00.htm
If you read the damn thing, you find that the filters weren't tested, but
rather the reviewers felt it was a priori reasonable to give them all
comparable good performance marks.
2) A consistent testing methodology that can be applied to several products
of different types does not exist.
3) Any test represents a point estimate at some point in time and at some
point in the network. This is quickly outdated as spammers evolve fast
(not that fast, but in a matter of months).
4) There are billions of spam mails circulating each day, how does one
generalize from a test involving a few thousand mails in a lab?
5) Nobody agrees on just what spam is, making 1-4 above irrelevant to some.
2. The vendor claims are true, but we're a bunch of perfectionists searching
for infinite resolution of Pi.
Some of us are perfectionists ;-) However, I think the main issue is that
the spam umbrella contains a large number of different subproblems for
different people, and solving one subproblem doesn't solve them all.
There's also no static spam problem, it evolves. So a formerly successful
solution or assumption can become outdated or useless. The virus industry
is a good example.
3. The filtering paradigm is a non-solution due to increased use of
recipient resources (bandwith, storage, processing, end-user time, etc)
I think that's true and false. It is a solution for end-users with
personal computers and fat enough pipes to get the input.
It is only an ingredient for middle-men or providers.
I also think that the vast majority of users world wide don't have
effective filtering set up, even though a small subset of power users
already use systems which can filter everything out that you throw at
it. So some people don't actually have a spam problem any more, but
they are in the minority.
I'm jaded here because I just don't get any spam at all. To me, this group
is firmly stuck in #2. But I'd love to hear everyone's feedback/experience
on "Why are we still here?"
I'm firmly a proponent of Bayesian end-user filtering. I view the
user problem as an information overload problem, which will only grow over
time. This overload will exist even if spam (ie unsolicited bulk mail
using stolen resources) was reduced to irrelevance.
For example, postal services allow bulk mail for an appropriate fee,
so there's no spamming in the postal system as such. But it's still a
disgusting practice from my point of view, as I have to accept, sift
and throw away legitimate, paid for garbage in shrill colours and big
lettering.
But rather than bore everyone with one sided arguments, I'm here to
learn more about the non-filtering aspects and help criticise them.
--
Laird Breyer.
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg