SpamCombat - part I

by Andrew Cooper

Foreword

Recently, our site was contacted by a representative of G-Lock Software Company, who asked us to review one of the company's products, SpamCombat - a program aimed to fight the spam clogging people's inboxes. Our editor had expressed interest in it, so he diligently relegated this job to me.

I was quite enthusiastic about the assignment and especially wanted to see how SpamCombat compared with the program that was in our previous antispam review.

I went to the developer's website, read some informative docs, and downloaded a 4-megabyte installation package from the server.

Installation and initial settings

After a brisk installation, the program offered to configure my mailbox settings to allow me to retrieve mail from multiple email accounts I use in everyday communications. I agreed to that "I'll make him an offer he can't refuse" question and gave the program all my account details (passwords, logins, and POP3 server names).



Please note that if you want to evaluate the effectiveness of your antispam software, you should scan your incoming mail with it before going to your regular mail checker. This will allow the antispam program (in this case, SpamCombat) to identify and remove spam stored remotely on your mail server before it reaches the local mailbox folder of your mail client. Thus, you will save time, disk space, and money that you would otherwise imprudently spend to pay for the extra bytes of downloaded data. Also, do not forget to check that you have correctly specified your email account settings. SpamCombat has such an option.

Back to the program - on the "Advanced" tab shown below, I have checked all possible mail filtering techniques offered by the program (highlighted only if "Spam Trap account" is enabled). Here, the program would offer to automatically delete mail messages it regards as spam, but be advised not to select this item when you haven't educated the spam filters adequately. (I did the contrary, as I didn't know how to use the program in the beginning.)

Filtering options

The program offers the following spam filtering techniques:

• Complex filter is a set of sophisticated algorithms employing a combination of various mail detection approaches to separate spam mail from legitimate email.
  
  
• Blacklists and whitelists are given much attention in our last article on spam.
  
  
• HTML Validator is a tool that looks into the contents of the HTML code (the basic language used for creating webpage documents) embedded in an email's body. It aims to identify links pointing to known spam resources on the Web. Once it determines that the code within the scrutinized email sample has links to definite spam sites or relates to products regularly advertised in spam, the filter flags the mail as spam and applies a deletion policy chosen by the user.
  
  
• DNSBL filter is a filtering module that collects data on spam from known antispam databases. Many antispam sites on the Internet share information about existing spam sources, and this filter processes the database entries to verify whether the mail in question actually originates in these places. If the filter confirms that the mail is indeed a spam, it won't let the mail through.
  
  
• Bayesian filter is the most powerful and advanced tool built into the program. It is a content analyzer. By employing a mathematical probability theory ability theory, the filter evaluates whether text given in a message exhibits the past traits of spam. Here's a simple example: If the filter knows that the text "Hello, you have been chosen to receive a car worth $80.000,00 today. To claim your prize NOW, click here www.justanexample.com" is a definite spam, it evaluates the chances that a similar but slightly altered message - "Hello, you have won a yacht worth $90.000,00 today. To claim your prize N O W, click here www.justanotherexample.com" - is also spam. It is a guessing game. You can customize the filter and set various levels of thresholds and tolerances. The Bayesian filtering instrument is a self-learning system that the user can educate to his or her own taste in order to increase the accuracy of the filtering process. An option to automatically delete spam from the user's mail server can be turned on after a certain period of time has passed and the Bayesian filter is educated enough to deliver a minimal false-positives ratio.

Continue to Part II