Home Original page

Graham's spam filter

Heiko Wundram hewu5001 at stud.uni-saarland.de
Thu Aug 22 23:52:44 EDT 2002 
Am Fre, 2002-08-23 um 02.44 schrieb Erik Max Francis:
> The client/server model also brings up another obvious issue:  Ask you
> what your roommates think about the potentiality for you to be able to
> read all of their email.

Err... I'm not trying to read their mail; that's why I propose only
keeping the spam data on the central server. What each person considers
spam doesn't necessarily vary very much; it basically is just an eMail
that does unsolicited advertising/unsolicited sending of
dialer-programs/unsolicited sending of virii.

And if I wanted to read their mail, I could do it much more easily
anyway, just installing a packet-sniffer on my PC (as the routers here
are so ill configured that every port on them is a monitor port, but I
can't change that, as I don't have admin-privs on them...).

But still, I think keeping a central database of SPAM signatures is
really the best way to start this, as the problem is that most users
don't really know how to configure something like this unless you give
them a course in configuring the system.

The problem isn't necessarily the spam itself, but as I have noted the
unsuspectable user. And I don't have the time nor the will to train them
on what to consider spam/virii/etc. and what not! Consider that I am
basically trying to write a program in my free time that will easen my
job, and their lives, as I don't have to go running around anymore
removing dialers or virii (hopefully).

Ask yourself what is worth more: either having a scanner that _may_
reveal a little of personal information (when the person using it marks
a legitimate mail as spam, and thus basically inserting the mail into
the bad corpus), or a system that is worthless, as most users don't take
the time/will to configure it...

Oh, and by the way: The system as it is doesn't store the messages, it
only stores the word-count in a database, so there isn't any realistic
chance that you will be able to reconstruct a message after it has been
inserted into the database.

Yours,

	Heiko Wundram
	Netzwart Wohnheim-D
	Zimmer 2206 - Universität 18 - Saarbrücken
More information about the Python-list mailing list