Opinion » Cybernaut


Going spam free



A recent industry study predicted that in the next three years, more than 60 per cent of all e-mails sent via Internet will be spam. Looking into my own Hotmail account, the number is closer to 90 per cent, but then I’ve made the mistake of buying tickets online from Ticketmaster, bidding on items on eBay, purchasing books from Amazon.com, signing on to mailing lists, registering for free online magazines, and booking at least three flights through different online services.

I put my name and address out there, and the chickens are coming home to roost – because I’ve spent money online before, the spammers are hoping that I’ll do it again. In the high-stakes carnival of online marketing, I’m what they would call a complete cyber-rube.

There are ways to cut down on your spam, however. Many of the services cost money, generally on a monthly basis. These scour through messages looking for giveaways in the Subject window and message text, as well as suspicious mailing addresses. As a result they have to be constantly updated to keep up with the spammers.

An interesting new service recommended by PC World Magazine (www.pcworld.com) to Outlook Express users, is SpamBayes.

Based on a theory espoused by an 18 th century mathematician by the name of Thomas Bayes, the program doesn’t accept or reject e-mail messages based on keys, but rather based on mathematical probability.

Each incoming message is scanned a number of different ways and assigned a number between 1 and 100. If a message scores a 90 or higher, it is automatically moved to a spam folder. If a message scores a 15 or lower, it’s moved into your inbox. All other messages are moved into a third folder where they can be sorted once again.

SpamBayes learns as it goes, getting to know what spams look like over time.

According to the author of the PC World report, the program is about 99 per cent effective, letting through the occasional spam. It has never done the opposite, channeling a genuine e-mail into a spam folder.

Based on open source codes, you can find SpamBayes at http://starship.python.net/crew/mhammond/spambayes/.

SAproxy by Stata Labs is also a good free choice, after coming out on top in Comsumer Reports’ latest antispam software tests, where it fared better in tests than no fewer than eight commercial anti-spam programs.

If you’re desperate and willing to pay for spam filtering, a good option might be Sunbelt Software’s IhateSpam (www.sunbelt-software.com), which is available for $30.

P2P or not P2P?

Sick of getting dragged through the mud, the Peer to Peer giants formed an alliance last week to challenge the recent spate of lawsuits against file swappers using their services.