ietf
[Top] [All Lists]

Re: spam

2003-05-27 12:20:14
Paul writes:

actually, i can speak for myself, there's no need
to interpret for me.

You may not be saying what you are thinking, however.

i'm calling you a troll because you're urging
people to pay no attention to the costs, to just
hit delete, and to avoid filtering since you claim
that the costs of false positives are too high.

You're calling me a troll because you disagree with me, and you hope to make
your opinion seem more objectively valid than mine by discrediting me
personally.  It's a poor debate technique, and not very persuasive to those
who are aware of the fallacy it contains.

those are the three key issues that every
spammer-in-troll raises whenever they join public
debates on the issue of spam.

I wouldn't know, as I don't recall having (knowingly) talked to spammers
directly.

usually it takes a week, but you did it all in
36 hours.  congrats?

I'm afraid I don't understand.

you act as if headroom would have no value.

It depends on the SD of the traffic and the desired uptime.  Yes, as a
matter of fact, I do know queuing theory, although it has been a few decades
since I last used it to any extent.

maybe that's true for you but again i'll suggest
that you avoid trying to interpret for me unless you're
offering to pay me $552 per month in perpetuity.

Since you've interpreted for me by making assumptions about my costs for
false positives and use of the delete key, I'm surprised that you object to
any assumptions I make concerning your costs.

But I'm speaking of a general case, not your case or mine.

in order to get 20:1 rejection rates at the smtp level,
i have to analyze the stuff.

How do you analyze stuff you reject?  Isn't the whole idea of rejection to
avoid processing?  And doesn't analysis require processing?

since the major cost is in receiving it, the minor
inconvenience of storing and analyzing it is pretty low.

But if you are storing and analyzing it, you're still paying the major cost
of receiving it.

the better my analysis, the less i receive ...

How do you avoid receiving what you don't want without any analysis or
storage?





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