Download: Spamihilator and K9
It's been over a week since I posted last, and I promised myself no matter what I'd be doing at least two updates per week. So I'll make up last week by posting a few more articles over the next few days. This first article is about preventing spam for home users.
I've been testing freeware anti-spam programs for quite some time. These days many Internet service providers (ISPs) offer email spam filtering. Typically the spam arrives in your inbox with the subject line marked. So instead of seeing "v1agra" you might see "***SUSPECTED SPAM*** v1agra" or something similar as the subject.
If your ISP does not offer a spam filter, you don't need to spend any money to filter out all that junk email! There are two good spam filters that I recommend.
K9
K9 is a simple but powerful anti-spam application that uses Bayesian analysis to determine good vs spam email.
One of the advantages of using this type of spam filter is that the more emails you get, the better it gets at predicting the spam. In fact it learns what you consider to be spam. This is an advantage over your ISP's spam filter, which always uses the same rule set and may continually let the same spam through or mark the same good emails as spam. K9 will not produce any of these false positives if you train it correctly.
K9 is a small download and automatically works with Outlook or Outlook Express. If you use another email client such as Eudora or Thunderbird, you will have to manually configure them (although Thunderbird has a built-in spam filter). As with any anti-spam software, you should create a SPAM folder in your email client to send messages that K9 marks as spam. This way, you can occasionally check the folder to see if it is targetting emails correctly. I found K9 to be the most accurate of all the spam filters I used (98.5% correctly identified on a regular basis) however as it's learning list got large it sometimes slowed down noticeably. However my computer wasn't particularly fast anyhow, newer computers may not slow at all.
K9 only works with POP3 mail; it will not work with web mail such as Hotmail, Yahoo mail, nor will it work with secure email that uses SSL, such as MSN.
Spamihilator
Spamihilator came highly recommended, and I was impressed by its polished interface. Spamihilator (a bit of a mouthful to pronounce) uses the same Bayesian technique as K9. It uses white lists and black lists so you can identify specific good and bad email addresses. As well it has a newsgroups feature which checks the recipient address instead of the from address, because some mailing lists use the same recipient address. This handy feature makes sure you get all your mailing list emails - no matter what the subject line is - without having to teach Spamihilator about each one.
Spamihilator also goes a step beyond K9 by employing a plugin system. This means that in addition to the Bayesian filter, other custom filters can be used. It comes with a few, such as an image filter and a file extension filter. The one main difference is that Spamihilator does not mark spam emails then send them into your email client, it actually deletes them before they get there. Fortunately it has its own recycle bin that keeps deleted emails around for a while so you can check it periodically.
Spamihilator supports IMAP as well as POP3, and also works with any email client, as well as secure connections over SSL. It also has a useful set up wizard that guides you through the setup. All in all a very configurable, excellent spam filter.
Conclusion
Remember that these spam filters use Bayesian filtering which takes a little bit of effort to train, and you should see good results in a week or two. While these programs are great for the home user, most corporations that have their own mail server will use a server-based anti-spam solution that can be controlled centrally by a network administrator.
QUICK DEFINITION ABOUT "SPAM" - Unsolicited email that was sent to you and thousands or millions of other people, also known as junk email.
I've been testing freeware anti-spam programs for quite some time. These days many Internet service providers (ISPs) offer email spam filtering. Typically the spam arrives in your inbox with the subject line marked. So instead of seeing "v1agra" you might see "***SUSPECTED SPAM*** v1agra" or something similar as the subject.
If your ISP does not offer a spam filter, you don't need to spend any money to filter out all that junk email! There are two good spam filters that I recommend.
K9
K9 is a simple but powerful anti-spam application that uses Bayesian analysis to determine good vs spam email.
QUICK DEFINITION ABOUT "BAYESIAN" - Bayesian refers to the Bayes theorem statistical analysis, which calculates probabilities of an email being spam, based on a variety of words and other techniques. Simply put, if the subject contains the word "viagra", there is a good chance of it being spam.
One of the advantages of using this type of spam filter is that the more emails you get, the better it gets at predicting the spam. In fact it learns what you consider to be spam. This is an advantage over your ISP's spam filter, which always uses the same rule set and may continually let the same spam through or mark the same good emails as spam. K9 will not produce any of these false positives if you train it correctly.
K9 is a small download and automatically works with Outlook or Outlook Express. If you use another email client such as Eudora or Thunderbird, you will have to manually configure them (although Thunderbird has a built-in spam filter). As with any anti-spam software, you should create a SPAM folder in your email client to send messages that K9 marks as spam. This way, you can occasionally check the folder to see if it is targetting emails correctly. I found K9 to be the most accurate of all the spam filters I used (98.5% correctly identified on a regular basis) however as it's learning list got large it sometimes slowed down noticeably. However my computer wasn't particularly fast anyhow, newer computers may not slow at all.
K9 only works with POP3 mail; it will not work with web mail such as Hotmail, Yahoo mail, nor will it work with secure email that uses SSL, such as MSN.
Spamihilator
Spamihilator came highly recommended, and I was impressed by its polished interface. Spamihilator (a bit of a mouthful to pronounce) uses the same Bayesian technique as K9. It uses white lists and black lists so you can identify specific good and bad email addresses. As well it has a newsgroups feature which checks the recipient address instead of the from address, because some mailing lists use the same recipient address. This handy feature makes sure you get all your mailing list emails - no matter what the subject line is - without having to teach Spamihilator about each one.
Spamihilator also goes a step beyond K9 by employing a plugin system. This means that in addition to the Bayesian filter, other custom filters can be used. It comes with a few, such as an image filter and a file extension filter. The one main difference is that Spamihilator does not mark spam emails then send them into your email client, it actually deletes them before they get there. Fortunately it has its own recycle bin that keeps deleted emails around for a while so you can check it periodically.
Spamihilator supports IMAP as well as POP3, and also works with any email client, as well as secure connections over SSL. It also has a useful set up wizard that guides you through the setup. All in all a very configurable, excellent spam filter.
Conclusion
Remember that these spam filters use Bayesian filtering which takes a little bit of effort to train, and you should see good results in a week or two. While these programs are great for the home user, most corporations that have their own mail server will use a server-based anti-spam solution that can be controlled centrally by a network administrator.
posted by Peter at
9:12 PM
