In My Experience... | Triple Spam Filtering doesn't work. (1282)

My Flash app: Part 2 « prev
I want a console better than the PSX. » next

Triple Spam Filtering doesn't work.
[ Posted by Dan on December 02, 2003 | 4 Comments ]

This post is the virtual version of me whining about something. Disregard as you see fit.

Anyway, this site is hosted at Pair.com, who have been a great host for the 2+ years I have been here, and one of the great things I can do is run my own CGI's and other apps on the host (assuming I don't do bad things like gobble up ton's of processing cycles). One of the things I do is run my email thru SpamAssasin (which can be a processor gobbler) when I connect with my mail client (ie, Apple's Mail.app client).

In Mail.app I have spam filtering turned on, and I have a static "rule" set up that catches spam as well. In spite of this triple filtering, about 20 messages a day, that I consider spam, are getting thru, and almost all of them are for drugs (ie, drugs that either stiffen certain things or make other stuff less aware of it's surroundings).

I think one of the things I don't like about the spam I get is how insulting it can be.

Dan, the Black List given by Spamcop.net stops 99% of the spam I receive. Bayesian filters stop the rest. Using multiple methods is the key. You can also add custom header filters (static rules as you call them) in your MTA configuration. You can see my MTA blocking results in real time on http://lashampoo.net/unix/SpamFilters

-Posted by Mike on December 3, 2003 04:22 AM

I'm using the SpamCop and SpamHaus lists on the server with SpamAssasin, which sets the email subject of suspected spam to specific string, and then Mail.app goes thru everything and trashes anything it thinks is spam and anything that has that specific string in the email subject as written by SpamAssasin. After all of that, my filter/rule takes a crack at it all.

I'm catching at least 90% of the spam I receive, but when 100+ spams arrive per day, 10 (or sometimes less than that) make it thru, which isn't a big deal really, but after putting in the effort to stop as much spam as I can, it's dissapointing to not acheive 99%.

-Posted by Dan on December 3, 2003 09:41 AM

It's not 99% but 99.5% that SpamCop catch. But that's true that I did cut my last email address too much spammed because too much present on web pages. As long you can find your email address on the net, Dan, you will be elligible to receive spam. There's some tricks on my website to mask your email address on webpages. This is the first part of the battle. But for your actual email address it's too late, you have to cut it and create another one.

Here's the plan:

Choose a non trivial address (no webmaster@yoursite.com or info@yoursite.com)

Never publish this email address on a web page "as it is" but chiffer it or use form to contact you. Be careful of newsgroups, forums, comments, mailing list archives that leave your email address lisible...

Use SpamCop and SpamAssassin on your server and bayesian filters email clients

Never enable the "Show images in HTML email" on your email client or click on links or ask to "remove my email address". Useally this is a trap, used by spammer, to validate your email address.

-Posted by Mike on December 3, 2003 10:32 AM

Just using AOL communicator and nothing else, I'm down to maybe 8 or 10 spams a day that aren't marked as such, and they're mostly ones which are using newer techniques, so they'll get filtered out as I continue telling the software that they're spam (only to be replaced by others using different techniques no doubt)... but 8 or 10 is a biig difference from 200-300.

Cutting all my email forwards helped a lot.. since of course I'd get the same spam in 2 or 3 email accounts. Just means I have to check multiple accounts, which isn't a hardship.

-Posted by JC on December 3, 2003 04:17 PM

Comment posting has been turned off because I don't have enough time and will to deal with the constant comment spamming. I'm very sorry and will fix this sometime soon (soon = before 2004 ends).