Spam, spam, spam, spam–how I got less of it

Someday we’ll get a spam filter at work, and the day can’t come soon enough. On Wednesday I got fed up with getting 8-12 messages a day from “Fulfillment Center”–I was much more irate than usual today–so I took a desperate measure.
I dug up the abandoned freeware Windows tool Bounce Spam Mail (search for it with your favorite search engine; there’s no official homepage). As spam came in, I pasted its headers into Bounce Spam Mail and sent back a bounce. Sending back bounces hours later has never been effective for me, but in these instances, I only sent bounces when I was sitting right there when the mail came in. And it seems to have helped a little.

Ultimately, it’s better to have SpamAssassin on the mail server, or install POPFile. I’m stuck with Outlook connected to an Exchange server in corporate workgroup mode at work, and the last I’d checked you couldn’t use POPFile that way. But checking last night, now there’s a way you can, by installing POPFile followed by an add-on called Outclass.

My fake bounces aren’t ideal, but at least they seem to be better than nothing. But I’ll be installing POPFile and Outclass really soon.

6 thoughts on “Spam, spam, spam, spam–how I got less of it

  • March 13, 2003 at 5:00 pm
    Permalink

    I installed them, and they work like any other Bayesian filter: At first they do nothing. Then they guess wrong for the first dozen or so messages that come in. They get smarter with time.

    The nice thing about POPFile is that it’ll also sort your messages for you, so automated messages can go into their own folder if you want, and not clutter your inbox (or be cluttered out by other stuff). Give it enough time, and it ought to even be able to figure out the difference between work and personal mail.

    You definitely want this.

  • March 13, 2003 at 5:51 pm
    Permalink

    comment
    This has nothing to do with spam (even though getting mail about increasing penis size or working from home blah blah blah, is f**ckn annoying!)

    I did a stupid thing last night. A friend thought I might need to upgrade my Windows 93SE to Windows ME, but didnt get me an upgrade version, but the full version. Through the installation process, I did not copy the unistall files to my hard drive as I didnt think I would need to uninstall, plus I wanted to get rid of all elements of any old system files, as they were the ones causing me problems (errors in cd-rom line 2, 3 4 etc)…
    Now, my computer stays in dos mode, and says that windows cannot run on ms dos version 8.00!?
    Please help!? I need to wipe my hard drive and start again. How do I do that?

    PLease help

    Kiki

  • March 13, 2003 at 9:31 pm
    Permalink

    I don’t get any spam at work; guess it’s a combination of our MTA servers doing a good job and the fact that I don’t do anything from work that clues anyone outside in to my work email address.

    At home I run my own email server and also pull mail from a pop server (with fetchmail) that gets most of the spam that I receive. I was running SpamAssassin, which I understand can be very effective. It worked moderately well, but that’s it; possibly because I didn’t add or pay for use of any spam list services. I got some spam still coming through, cranked up the sensitivity of the filter, and started getting false positives.

    I figured that if I have to keep fine tuning the program, it’s no better than continually adding to procmail recipes. I decided to install SpamBayes ( http://spambayes.sourceforge.net/ ) on my server. You install it, and set up a cron job with mail folders defined as either spam or “ham” (non-spam). The cron job builds the SpamBayes database so that it learns to recognize the difference in spam and “ham”. You just have to put spam in a folder designated as being for spam and leave it there long enough for the cron job to run and train SpamBayes.

    Similar to SpamAssassin, SpamBayes writes a header based on what it sees in the email, and I use procmail to filter mail as spam or not. Like Dave says, bayesian based programs take time to learn, and at first do nothing useful. “Unfortunately”, about the time that I installed SpamBayes, the ISP where I have my POP mailbox installed spam filtering, depriving me of much of the spam to use in training SpamBayes.

    However, it’s gotten enough training now ( a week or so later, with some old spam rounded up and used ) that it’s starting to recognize spam accurately.

    Time will tell, but I’m optimistic that the program will achieve what I want. For a while I classify mail as spam or not. Then it does for me what I would do, without the need for intervention from me.

Comments are closed.

%d bloggers like this:
WordPress Appliance - Powered by TurnKey Linux