Hi Thager, your question on Brightmail has already been answered by dalek in your thread about it.
There are several approaches to spam filtering:
1. Filter for illegal words in the header and/or content. Things like "penis", "viagra", "loan" and so on.
2. Use a Bayesian filter. This rates individual words and uses some sort of algorithm to calculate a value, then accept or reject an item. At its crudest this would be to just divide by the total number of words to give you a "raw score" You might have seen really weird spam mails with apparently random sentences.............that is to cheat the Bayesian scoring system ;)
3. Use a Blacklist/Whitelist system. This blocks IP addresses that are associated with spam. Brightmail works like this.
4. Block ISPs and regions. Like if you never do business with Nigeria, you don't want e-mail from them do you? :eek:
Hope that helps
