|
Dear Spam-haters (on Linux RH8),
At first I parsed headers
manually, e-mailing authorities as best I could.
Very time consuming and
pointless.
I found an article on writing a
'custom action' for the rules in my M$ based client. I wrote rules and put
bad addresses on a list, but the list kept growing and required daily
maintenance.
Time consuming and
pointless.
I found SpamCop. Automatic
headerparsing and e-mail generation. Cool! I used SpamCop for a long
time, but the true benefit of SpamCop is realized only if you are a
subscriber. If you are, then you get your mail through them and they block
all spam reported to them. If you keep using your old method of getting
mail it doesn't really have any noticeable effect.
Time consuming and pointless
unless you subscribe.
Finally, I saw an article in
Linux Magazine on the language Ruby, and after playing with it a while I wrote a
'despam' utility which runs on a Linux box.
The theory I went with is 'all
e-mail is spam'. Then there are those rare exceptions when someone you
actually know writes to you and you want it, so that gets passed through.
They get put on a 'friends list'. All others are moved to a 'soon to be
deleted' folder. There are problems with
both the theory and the program, but it does help a lot.
If you've been on the Psyche
list for a while you may have seen one of the problems. I joined the list
back in October and suddenly I started getting these (ranging from polite to
hysterical) replies that I was spamming the list. Part of the program is a
(now optional) reply to the (usually fake) From: address. I did it without
thinking and it clashed with my system and so a lot of people got replies that
were simple list submitters (some call it a 'feature'). I think if I had
selected 'digest' format and pre-added the 'psyche-list-request@redhat.com'
name as a friend it probably would have been 'ok'.
I wanted a system that was low
maintenance so I wanted some automated features. The (now optional) reply
informed you that if you forwarded this to 'me@myserver' it would pass because of a
signature in the message and I would get an e-mail with subject 'REQUEST FRIEND
STATUS'. If I replied in the affirmative then the program would add them
to the list and they could send me mail. If a spammer actually put their
real e-mail address in the headers and they got the message and forwarded it to
me, I would just delete the message. If it became a problem because they
had the signature, I just change the signature in the program. If they got
smart, they could automate replies and spam me anyway, but this is war!
Isn't it?
Another problem is that my
program pops the server at 1 minute intervals. My e-mail client pops the
server at 15 minute intervals. If my client pops the server in the period
after the spam is sent but before the despammer cleans the
account then I get a little spam anyway. It needs to be a proxy
server. Anyway, it's been turned off
for the last two months because I've been looking for a job and I just didn't
want to mess with possible job offers being deleted. And I haven't
been advancing the work on the program during this time.
There is probably an
existing program like 'fetchmail' or 'SpamAssassin' (or maybe even qPopper)
that could be set to this mode but I just don't know of it.
Anyone...?
As someone mentioned earlier,
the only effective way to stop spam is to block it at a server. This is my
attempt. I don't think there is really a good solution yet. It
should be illegal, like littering, and spammers should be fined, but that would
only work for US based companies. What about countries that hate us?
They probably pay people to generate fake traffic. We need some real
creative thinking and a coordinated effort.
Thanks,
Greg
|
