Filters still no match for non-English spam

Dom · Jun 24, 2003, 07:41 PM

LONDON (Reuters) - What's "spam" in Spanish? Or what about Chinese?

Software technicians are grappling with these questions as they toil on new filters designed to intercept the burgeoning flow of unsolicited e-mail -- known the world over as "spam" -- that land in e-mail in-boxes daily.

The latest spam-blocking software on the market is stopping cold billions of unwanted messages daily, offering from cheap home loans to anatomy-enhancing toolkits. But most of the spam-fighting technology is designed to detect only English-language messages, experts say.

"Some languages, like Japanese, it's not easy at all to detect if it's spam," said Gene Hodges, president of Santa Clara, California-based Network Associates.

Network Associates, one of a score of vendors that sell spam-detection filters to corporations and individual computer users, is aiming to introduce over the next six to twelve months software that will pick off spam written in Chinese, Japanese, French, German and Dutch, to name a few.

BILLIONS AND BILLIONS SERVED

The world's largest e-mail providers Yahoo, AOL Time Warner's America Online unit and Microsoft already block several billion spam messages daily. And politicians in the U.S. and Europe are drafting laws to criminalise the activity.

But computer security experts warn such measures will have a limited short-term impact, particularly as spam artists continue to disguise their location and identity to fool the software filters and obscure their legal whereabouts.

Hodges estimated that in the past six months the percentage of spam written in English has declined from 95 percent of all spam to 80 percent.

As spam levels continue to rise globally, the number of get-rich-quick schemes written in Chinese characters, for example, are becoming increasingly common, he said.

Typical spam detection technology is based on a complex algorithm that scores the likeliness of an e-mail message's authenticity by scanning the contents of the subject line and the message itself, plus the e-mail address of the sender.

The filter works by sniffing out certain suspect phrases such as words pertaining to finance or pornography and labelling it as "spam".

Tweaking the system for messages written in Romance languages is relatively straightforward. Spam written in Asian languages, with vastly different character sets and sentence structure, is much more time-consuming.

"With spammers spreading as fast as they have to non-English speaking regions, the race is on," Hodges said.

_____________________
Source: Reuters

The_Neon_Cowboy · Jun 24, 2003, 10:52 PM

just nuke beigejing , parts of russia, and parts of europe thats where 95% of them spam is really coming from.. out of the thousands of peices that i have backtracked...

Jun 24, 2003, 07:41 PM	#1
Dom DriverHeaven Extreme Member Join Date: Jun 2002 Posts: 12,940 Rep Power: 0	Filters still no match for non-English spam LONDON (Reuters) - What's "spam" in Spanish? Or what about Chinese? Software technicians are grappling with these questions as they toil on new filters designed to intercept the burgeoning flow of unsolicited e-mail -- known the world over as "spam" -- that land in e-mail in-boxes daily. The latest spam-blocking software on the market is stopping cold billions of unwanted messages daily, offering from cheap home loans to anatomy-enhancing toolkits. But most of the spam-fighting technology is designed to detect only English-language messages, experts say. "Some languages, like Japanese, it's not easy at all to detect if it's spam," said Gene Hodges, president of Santa Clara, California-based Network Associates. Network Associates, one of a score of vendors that sell spam-detection filters to corporations and individual computer users, is aiming to introduce over the next six to twelve months software that will pick off spam written in Chinese, Japanese, French, German and Dutch, to name a few. BILLIONS AND BILLIONS SERVED The world's largest e-mail providers Yahoo, AOL Time Warner's America Online unit and Microsoft already block several billion spam messages daily. And politicians in the U.S. and Europe are drafting laws to criminalise the activity. But computer security experts warn such measures will have a limited short-term impact, particularly as spam artists continue to disguise their location and identity to fool the software filters and obscure their legal whereabouts. Hodges estimated that in the past six months the percentage of spam written in English has declined from 95 percent of all spam to 80 percent. As spam levels continue to rise globally, the number of get-rich-quick schemes written in Chinese characters, for example, are becoming increasingly common, he said. Typical spam detection technology is based on a complex algorithm that scores the likeliness of an e-mail message's authenticity by scanning the contents of the subject line and the message itself, plus the e-mail address of the sender. The filter works by sniffing out certain suspect phrases such as words pertaining to finance or pornography and labelling it as "spam". Tweaking the system for messages written in Romance languages is relatively straightforward. Spam written in Asian languages, with vastly different character sets and sentence structure, is much more time-consuming. "With spammers spreading as fast as they have to non-English speaking regions, the race is on," Hodges said. _____________________ Source: Reuters

Jun 24, 2003, 10:52 PM	#2
The_Neon_Cowboy HardwareHeaven Extreme Member Join Date: Dec 2002 Location: U.S.A. Posts: 16,009 Rep Power: 90 System Specs	just nuke beigejing , parts of russia, and parts of europe thats where 95% of them spam is really coming from.. out of the thousands of peices that i have backtracked... __________________