Pron Maps

Jun. 10th, 2004 03:14 pm
ltmurnau: (Default)
[personal profile] ltmurnau
I don't know about you, but lately I've been getting a lot of spam. None of the messages are large, but I seem to get about one megabyte a month of the stuff.

Originally it was lots of stuff offering me all kinds of prescription drugs, mostly pain killers or stimulants, but recently the balance has tilted way in favour of pornography. I had some filters in place with my ISP, but these seem to be defeated through two methods:

I. Creative misspelling. The most recent e-mail I got featured:

"Eexemtly Huge cawks! Massive cawks not seen anywrehe else!

See this slluts take thsee massive cawks in every hole! Many Many mroe x-raited adevntures of mstoenr cawkks!!!

Cmoe cehck tsehe wild vixnes work teshe cawks hard!!!


These keyboard-mashing examples of someone (let's hope) figuratively typing with his dick would defeat any kind of automated spam filter that relies on proper spelling.

II. Random text, both in the subject line body of the message. The "cawk" introduction was followed by long passages grabbed at random from several unrelated documents:

Government documents: "Other Interesting Facts of Employees Rights Under the ADA"

Chinese history: "Problems produced by Hua Guofeng, the man who had succeeded Mao Zedong as..."

Something about the space program?: "James E. Webb took all responsibility for the tragic accident, but it was the hatch that had a design flaw that made it impossible for escape, " We've always known something like this was going to happen sooner or later, Who would have thought the first tragedy would be on the ground?" (Smith 7). Later Webb would decide to retire in October 1969.

And finally, an account of the life of Buddha and the history and basics of Buddhist practice that went on for several thousand words.

Very curious. I'm not sure how this method manages to foil filters.

Date: 2004-06-11 07:44 am (UTC)
From: [identity profile] cherdt.livejournal.com
The random text is to fool Bayesian filters, where certain spam keywords increase the probability that a message will get filtered (mortgage, free, webcam, viagra) but words that do not typically appear in spam (adage, triptych, infundibular) decrease the probability that a message will get filtered.

Bayesian filtering seemed like a great solution at first, because it would help prevent false-positives--legit messages that also contain spam keywords ("Hi Mom! Just wanted to let you now the bank approved our mortgage! Can I borrow that $10,000 now?"). But inserting the random text seems pretty good at fooling it.

However, I think someone could pretty easily solve that by analyzing legitimate text and comparing the word sequences to the random text (there are some programs that do such work for weird and fun purposes at Fun with Markov Chains). The random text seldom contains any verbs, from what I've seen, and it shouldn't be too hard for even a computer to see that the strings of words aren't coherent sentences.

Unfortunately, even though I'm a computer geek, I have no idea how to go about programming such a filter.

Date: 2004-06-11 08:51 am (UTC)
From: [identity profile] ltmurnau.livejournal.com
Thanks for filling me in! I did not know how filters worked, I thought they looked for certain spammy words and left it at that.

What amazes me is that spam is getting worse and worse. How can there be enough people out there horny enough AND poor enough AND dim enough AND zonked enough to consolidate their bills in a little blue pill that guarantees them a penis the size of a dinner ham and regular visits from a "cwm-sukkng tean dildoe maniakkk" to keep the whole scheme going?

Date: 2004-06-11 10:58 am (UTC)
From: [identity profile] cherdt.livejournal.com
I always used to think spam was sent by get-rich-quick schemers fed up with their lousy day jobs who were lured by the thought of Internet gold. How else could someone be so blind to the basic fact, that no one ever responds to a spam message?

And then I read this Wired article: Swollen Orders Show Spam's Allure

6000 orders for penis enlargement pills in 4 weeks. 2 bottles per order. $40 profit per bottle. That's $480,000 in a month.

How can I really be mad at spammers when they're really playing P.T. Barnum's part in this scenario? Instead I'm fed up with all the guys that are concerned their penis is too small. I want to say, "Hey, guys. That girl that dumped you? It wasn't because of your penis. It's because you're a goddamned idiot!"

Date: 2004-06-11 03:44 pm (UTC)
From: [identity profile] ltmurnau.livejournal.com
That's an amazing story. I wonder how many hundreds of thousands of spam e-mails were sent out to scrape up 6,000 idiots who had both a credit card and an Internet connection. But you do have a point: as long as spam works, we'll have spam.

Profile

ltmurnau: (Default)
ltmurnau

November 2024

S M T W T F S
     12
3456789
1011121314 1516
17181920212223
24252627282930

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Jul. 16th, 2025 05:29 pm
Powered by Dreamwidth Studios