NetGuide
search: 
you are here >

How To

The Guide

Home
Latest Issue
Email

How to kill spam

Out, damned spam!
How to kill spam with two free filters By Rose Vines

Most email sent each day (around 95%) is spam. So, what can you, poor beleaguered computer user, do to solve this dilemma? Reach for a spam blocker. There are dozens of these handy tools available and some of the best of the bunch are free.

Even if you are smart enough to avoid clicking or responding to the spam you receive, you still have to waste time dealing with the stuff. And because spammers try to rope you in by using deceptive subject lines and false sender addresses, figuring out what is and isn't spam often entails more than a quick glance and a press of the Delete key.

We're going to take a look at two spam solutions in this article: the SpamBayes Outlook Add-In and Thunderbird's Junk Mail Controls. The first works hand in glove with Microsoft Outlook (but not with Outlook Express); the second is a set of built-in spam-handling tools supplied with Mozilla Thunderbird, the companion email program for the Firefox browser.

 

Bayesian filtering
SpamBayes and Thunderbird both use a spam blocking technique called ‘Bayesian filtering’. In essence, Bayesian filtering assigns a ‘spamminess’ weighting to words that are more likely to appear in spam, and then assesses whether a particular email is likely to be spam or ham (not spam) based on the general spamminess of its content.

This approach is remarkably successful, but it relies on some initial effort on your part. You must teach the filter to distinguish between spam and ham by feeding it a bunch of messages and telling it which is spam, which is ham. This training phase helps the Bayesian filter adapt to your particular needs.

For example, ‘Viagra’ is generally a very spammy word; but if you happen to work for Pfizer, the company that makes Viagra, then your daily flow of business

email may well contain references to the drug. The training session allows for such differences and lets the Pfizer salesperson send and receive emails about Viagra, but still block emails that contain the word Viagra in combination with other spammy words such as ‘cheap’ and ‘Now!’.

 

Using SpamBayes
Microsoft Outlook, the big sister to Outlook Express, comes with a junk mail filter built in. It does a fair-to-middling job of screening out spam, but there are plenty of spam blockers that do a better job. SpamBayes Outlook Add-in is one of them.

SpamBayes has the twin virtues of blocking spam efficiently and costing nothing. It's also very easy to use, although you may find the download site somewhat confusing.

To use SpamBayes, you'll need a copy of Microsoft Outlook 2000 or later running on Windows 98 or later. Here’s how to install SpamBayes.

  1. Close Outlook if it's currently open.

  2. Visit spambayes.sourceforge.net/windows.html and click the Download The Installation Program link. You'll be presented with a list of download mirror sites. Click any of the Download links to download the program - the Australian link should provide the quickest download - and save the download to your desktop. We’ve also included it on this month’s CD for your convenience.

  3. Double-click the downloaded file to run the setup program.

  4. During setup, tick the Microsoft Outlook Addin box and leave the ‘Server/Proxy application’ box unticked, then let SpamBayes install to the default folder.

  5. Open Outlook. When you do so, the SpamBayes Configuration Wizard will open.

  6. You have a number of options for training SpamBayes to differentiate between spam and ham. By far the simplest method is simply to train as you go, so click ‘I Haven't Prepared For SpamBayes At All’ and click Next.

  7. Click ‘I Want To Continue Without Training’ and click Next.

  8. SpamBayes needs to monitor your incoming mail folders. Almost always that's the Inbox; if you use Outlook rules to sort mail automatically into other folders, click the Browse button and tick each of the relevant folders, then click OK and Next.

  9. SpamBayes sends email that is clearly spam to a spam folder, and suspect mail to an ‘unsure’ folder. You can name these folders as you like, then click Next and Finish.

 

Training SpamBayes
Without training, SpamBayes sticks all your email into the "unsure" folder. Teaching SpamBayes to become smart about spam and ham couldn't be simpler. Here’s how:

  1. Open the ‘unsure’ mail folder.

  2. Click each email in turn. If it's spam, click the Delete As Spam button, which SpamBayes has added to the Outlook toolbar. If it's ham, click Recover From Spam. To speed up the process, you can select multiple spam or multiple ham by holding down the Ctrl key while you click each item, and then click the Delete or Recover button as appropriate.

Repeat this training process each time you receive new email. After several days - depending on the volume of email you receive - you'll find SpamBayes becomes adept at doing the sorting for you. Continue to monitor the ‘unsure’ folder, and use the toolbar buttons to re-sort any miscategorised items in your Inbox and Junk E-mail folders.

 

Using Thunderbird
Thunderbird provides a couple of ways to help you sort spam from your wanted email. It starts by letting you give a free pass to all email from senders in your Address Book. This way, all you have to do to ensure email from your regular correspondents is never blocked, no matter what the content, is to add them to the Address Book. To filter the rest of your email, Thunderbird provides a Bayesian-based Adaptive Filter.

Here's how to set up spam filtering in Thunderbird.

  1. Click Junk Mail Controls from the Tools Menu.

  2. On the Settings tab, tick ‘Do Not Mark Messages As Junk If The Sender Is In Personal Address Book’. (Note: To add people you write to to your address book select Options from the Tools Menu, click Composition and tick ‘Automatically Add Outgoing E-mail Addresses To My Personal Address Book’.)

  3. Still on the Settings tab, tick the option ‘When Displaying HTML Messages Marked As Junk, Sanitize The HTML’. This is important, because it protects you from embedded links and other items in the mail which might install spyware on your computer or lead you to a phishing site on the Web.

  4. Adjust any of the other items on the Settings tab to suit your needs.

  5. Click the Adaptive Filter tab and tick Enable Adaptive Junk Mail Detection, then click OK.

 

Training Thunderbird
Training Thunderbird's adaptive filter is a simple matter:

  1. To identify spam, click the email in your Inbox - or any other folder - and then click the Junk icon on the toolbar or click in the Junk Status column. The email will be marked as spam and either moved to the Junk folder or deleted, depending on the option you chose when adjusting the Junk Mail Controls.

  2. To identify good email that Thunderbird has categorised as spam, open the Junk mail folder, click the email, and then click the ‘Not Junk’ icon in the toolbar (or click the ‘This Is Not Junk’ button at the top of the message preview window).

  3. Thunderbird also keeps an eye out for phishing scams - email designed to get you to divulge personal information. It places suspect phishing mail in your junk mail folder and marks it as a scam. If Thunderbird makes a mistake and marks legitimate mail as phishing, click the mail and then click the ‘Not A Scam’ button at the top of the preview window.

Annoyingly, Thunderbird fails to move mail you've identified as ‘Not Spam’ or ‘Not A Scam’ out of the Junk mail folder and back into your Inbox, so you'll need to do that manually by dragging the email into the appropriate folder.

As with any Bayesian filter, Thunderbird's spam filter will become better at its job as it trains. At first, you'll need to sort most of your mail, but within a couple of weeks Thunderbird should be blocking most of your spam for you. Remember to do a manual check of the contents of your Junk mail folder to ensure that no false positives have slipped through.

Â

Useful Web site
Wikipedia has an excellent series of articles on spam in all its forms, including ‘spim’ (spam delivered via instant messaging programs). You'll find it at
http://tinyurl.com/5zyxh

Â

Tips

Outlook Express
Looking for spam protection for Outlook Express? Try CloudMark Desktop (
www.cloudmark.com/desktop/?from=cnsmr-box) or Ella (www.openfieldsoftware.com). Neither is free, but each offers good protection.

Suite protection
Some Internet security suites, such as PC-cillin Internet Security 2006 (
www.trendmicro.co.nz) include spam protection. If you're looking for one-stop anti-virus, anti-spam and hacker protection, these suites are good value.

Don't reply!
Don't respond to spam. Ever. Not only does it encourage spammers; clicking a link in spam may also load spyware onto your computer. And don't bother clicking the ‘remove me from your mailing list’ links in spam; they don't work, and sometimes they're used to confirm your email address.

SpamBayes POP
If you don't use Microsoft Outlook, you can still use SpamBayes with any POP3 email account by installing the sb_server. Be warned, though: it's highly finicky to configure.

Beware the false positive
No spam blocker works perfectly. Sometimes spam will be classified as ham (a false negative) and it will end up in your inbox. That's a nuisance, but nothing major. Occasionally, though, ham will be classified as spam (a false positive) and that's a much greater worry because you could easily end up deleting legitimate email.

Dictionary spam
One of the ways spammers try to defeat Bayesian filters is to toss a whole pile of words with a low spamminess weighting into the subject line or content of their emails. The aim is to lower the overall spam score of the email. That's why you get emails with bizarre subjects like "hedgehog Gethsemane subsumed".

PRINT THISEMAIL THIS
Share this page - del.icio.us del.icio.us | digg digg | technorati technorati | reddit reddit
NetGuide Village
Click here to Register
Register now (it's completely FREE!) and unlock the NetGuide Village content!
Click here to Login
You'll be redirected to your villager profile

sponsored links