Spam is a daily hassle for anyone that relies on sifting through their inbox to keep track of things. I know that when I was at Yahoo, my inbox received probably a hundred emails a day, from mailing lists to emails from others in my group. Fortunately, Yahoo had some strong spam prevention measures. For others though, dealing with spam is a hassle that can't be remedied easily. I'll be checking to see if Agnitum's Spam Terrier plugin for Microsoft Outlook and Outlook Express can help the situation.
Spam Terrier installation is a trivial process and would be easy enough for my sister, who relies on Outlook throughout her work day, to install. It relies on the Bayesian filter which has been shown to adapt and teach itself similar to how many pieces of firewall software learn what ports to block or allow after a short period of time.
The Bayesian filter is based on probability algorithms that assess the odds of new messages having the same characteristics as past emails designated as spam. The more thoroughly you train Spam Terrier, the more accurate and tailored to your requirements the results will be.
Before I go any further, let me make it clear that Spam Terrier is a free plugin, with no catches. Although, you must register for a free key - but why must you need a key to use free software?
The main interface for Spam Terrier consists of a simple tool bar. When you first begin, it is a good idea to train it. The best way of doing this is to follow the train wizard which asks you to select a folder with spam/junk emails in it and it will process them.
Furthermore, you have the option of setting up a white list to keep email from various domains, email addresses, IPs or keywords safe as well as a black list to treat emails matching your criteria as spam. Feel like ignoring every email from your ex? The black list makes it very easy. In addition you may also customize the sensitivity of Spam Terrier yourself. The way the Bayesian filter works is that it gives emails a "rank" depending upon the type of words used in the email as well as their frequency of use, etcetera. It checks this against what it already knows from training and sorts messages as spam or probable spam depending on how high this rank is.
I had Spam Terrier process 600 emails in my inbox that I had previously gone through manually to clean out spam email messages to test if it would return any false positives. On the "normal" sensitivity setting, Spam Terrier returned 3 spam messages that were not spam. However I can understand why Spam Terrier flagged them as such. Taking the following email as an example, it was sent to a mailing list and used characters like the dollar sign and words like "contest", which are often used in real spam emails.
Fortunately, setting Spam Terrier straight is an easy process. You can either setup a white list rule so email from that person or mailing list is always safe or you could "Mark as Not Spam" on a per-email basis. I also subjected Spam Terrier to a folder of roughly 30 of those "new" spam messages that contain graphics and text designed to get past anti-spam measures. Spam Terrier was correctly able to identify each email as spam. Agnitum describes those types of emails a bit more on their website:
Spam Terrier successfully recognizes even the most sophisticated spam attacks â€“ messages that imitate regular emails to get past anti-spam software. These messages typically include embedded graphics and random words from the dictionary, which combine to make the message appear legitimate.
Overall, Agnitum's Spam Terrier struck me a simple yet effective method of spam protection for Outlook users. I don't use Outlook regularly, nor do I understand how people could willingly use Outlook as their primary email client when there are great alternatives such as Thunderbird. However, for people tied to Outlook through their company/work, Spam Terrier is definitely something to check out. Speaking of Thunderbird, there is a similar plugin called ThunderBayes that also utilizes the Bayesian filter and lets you flag emails as spam or ham.
Disclosure: I received 2 crisp C-notes for this review.