vetaltm Author
Joined: 05 Feb 2006 Posts: 748
|
Posted: Sat Oct 11, 2008 6:36 pm Post subject: |
|
|
The main objective of antispam program is saving your time, which is wasted when you are sorting the mail manually. You are looking at the mail headers, trying to guess if a message is good or bad and selecting the items to delete. It looks easy when you have 10-20 messages in the mailbox. But sorting out hundreds of messages, mostly spam, is much harder.
After a short period of additional training the plug-in can filter out more than 99% of spam automatically. It deletes from server the obvious spam classified by message headers, then downloads the rest of messages and classifies them by full content. The usual workflow is the following:
- The plug-in does its job - blocks spam and passes through the rest of messages.
- You are reading the received mail and training the plug-in on missed spam. It is not hard, because the trained plug-in can miss only 1-2 spam messages from 200-300. Thereby you can concentrate on your tasks instead of digging the heaps of spam every time on receiving new mail.
- When you have a time (e.g. once a day) it is possible to take a quick look at the Filtering log and see if the plug-in have done its work properly. The probability of false positives is very low, so in most cases there will be no mistakes. But it is possible to restore a mistakenly blocked message, even if it was blocked by header on the server.
Each approach has own pros and cons. When you preview the message headers before downloading, it is possible save 100 or more kilobytes of traffic, depending on the number of spam messages. But the plug-in also deletes spam by headers. By default it deletes from server 10-40% of spam, classified by sender IP, by black words in subject and by black rules. In most cases it is possible to customize the filtering rules and make the program to delete more than 60% of spam without downloading the message body. I suggest you to try this way, because in general reading good mails is more interesting than looking at spam headers |
|