On Sat, Mar 29, 2003 at 08:55:35PM +0000, Jon Kyme wrote:
I'm the naked man in a glass room - I have no way of knowing how many
recipients the message has. I can never know I've been spammed?
Next!
Really? I must say I have the exact opposite experience. Almost never
to I get a spam that I can't immediately identify that it was not written
personally for me. That alone of course doesn't prove that it's bulk
but it's a remarkably accurate indicator.
If you are building a bulk detector to prove bulk (for example to enforce
a legal regime) the next step is fairly obvious -- forward any messages
you get which _look_ to your fairly accurate intuition as bulk to a common
spot, where messages that meet similarity metrics can easily be pooled for
confirmation.
Of course, that's only if you want to do things based on individual action.
Clearly the mail servers for a large body of users can detect bulk the old
fashioned way, but getting multiple messages. Brightmail has hundreds of
thousands of spamtraps, and detects bulk that way (but also just uses the
old fashioned human inspection trick on messages that come to the traps too)
And as I have proposed, a set of MXs representing millions of users and pooling
data can easily detect bulk by simply looking at delivery volume from network
blocks, no need to even look at the content.
To my mind, trying to figure if a message was written for me, or is part of
a mass mailing, is remarkably easy and accurate. I would be curious to hear
if others have had trouble doing so as Mr. Kyme appears to report.
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg