Friday, 25 June 2004
Even Monty Python Would Have a Hard Time Making This Funny
I get spam, lots of it. Due to my youthful indiscretions involving where I stuck my email address, every Joe Penis-Enlarger and Jane Horny-Co-Ed now want to be my pals (or bedmates, it's all a bit confusing). Yes, I left my address unprotected, and as a result I'm supposed to Earn Easy Ca$$$h!!! This may or may not be related to helping out most of the population of Nigeria with URGENT BUSINESS ASSISTANCE to the order of ABOUT TWENTY MILLION UNITED STATES DOLLAR [sic] (or "$20 MILLION," in case you didn't understand ALL THE SHOUTING).
And, in the ultimate of insults, I am plagued by spam from vendors who want to sell me anti-spam software.
Yea, I say unto you: I have sinned. I am wracked with the guilt of - gasp! - having put my email address on the Web.
Back when I first got my domain, I was happily chugging along with my ISP-provided address. I was an early customer, and also the first "John" to think of asking for his name. So by the time I registered "johnkeller.com" in 1999, I already had years of history and many people with my old email address.
Why not, I thought, include a Web version my resume ("CV" for some) at my new domain? Nifty. And why not then post a very personal - and easy-to-remember - email address along with it? So I did.
And now, years after having abandoned my dialup account (it's sort of expensive to dial in to Minneapolis from Paris) and switching over to what was once my resume's dedicated email address, I'm suffering the consequences. This is, of course, long after having removed this said address from my Web pages. Ironically, the address currently posted (and has been for at least five years), gets almost zero spam messages. Sigh.
Fortunately, there are salves and balms for the pain, if no outright remedies. My main tool in the kit is SpamAssassin, which sits on my host's server and tags suspect messages for me. With the help of a technique called Bayesian filtering, SpamAssassin is very effective in evaluating the likelihood that a given message is spam. My email arrives as usual, with certain (e.g. "many") messages tagged as spam. And if SpamAssassin messes up, I can just feed the message back to it so that it learns better for next time.
There's always the potential for email tagged as spam that is, in fact, "ham." So I still have to slog through those damn messages - which number several hundred a week. In the future, when I'm being bombarded with 10,000 offers every minute for some futuristic version of V1@GRA, I'm sure that I'll look back on these halcyon days with fondness. But right now, "several hundred a week" is a lot.
SpamAssassin evaluates each message using rules (does it look like spam, taste like spam, smell like spam...). These rules are assigned point values, which add up to a "is it or isn't it" threshold. After some extensive tweaking, I've found a combination of rules and a rating threshold that avoid too many false positives or false negatives.
Well, I finally was able to go one step further. Yesterday, I added a mail-filtering rule (using another service that exists on my host's server) that applies a second limit. All spam with a rating above this second, slightly higher, threshold, is summarily dumped (well, I still get it for testing purposes, but that will soon change).
This effectively reduces the number of messages that I have to manually evaluate: Since setting the gears in motion, only four spam messages have gotten through in 24 hours. Forty seven others (47 OTHERS) hit a brick wall (the proverbial "/dev/null," in geek parlance) - or would have, were I not still in testing mode.
Before, it was annoying whenever I got a spam message. Now I say, "bring it on." It's still annoying, but also sort of fun - sort of like when young kids become annoying by never getting tired of playing the same, simple game.
So, I'm on my way to (mostly) spam-free email. My question is, where is our lawmakers' spine and why is it me who has to bear the onus of this whole process?
Maybe one of these days, I'll get an email promising me to be SPAM-FREE FOR THE REST OF YOUR LIFE if only I ***Clik heer***. Of course, it'll probably just get automatically filtered out.
Comments
Urgent message:
my name is Sven Johnson. I live in a little village in North Dakota. I have to eat lefsa and steak every day because my entire family died trying to milk the cows on our farm.
If you send me your account number, along with a urine sample, birth certificate, and strand of hair for the DNA test, I will send you 2 gazillion dollars because the I sued the cows and won a great inheritance that I cannot get due to my being in prison for treason. Send all this stuff to me soon, before the Gestapo gets you.
Falo (the girlie) Halo
Ha! Here's something that's hilarious on so many levels:
You can't see them now, but I had three - count 'em, three - "comment spam" comments hanging off this entry. The first ones I've ever gotten. Some script picked up this entry, out of everything I've written, to add its crap. Some post-modern confirmation of my existence?
I was almost - almost - tempted to keep them. But, nah.