Use of a spam signature database to identify offending material as it comes in, is only a start. Without fingerprint database analysis, lexical analysis and statistical analysis, augmented with the use of artificial intelligence techniques, Boolean logic and heuristics, no attempt to block spam will be wholly effective.
Fingerprinting will help identify all known spam strains and avoid falsely blocking an innocuous message. Lexical scrutiny will pick up key words and phrases in a message text in the context of the whole email. Use of artificial intelligence to develop a rules base of what an organization defines as spam, and statistical analysis to weight the overall probability that a message is spam, will further enhance effectiveness, it is claimed.
The prognosis from the Congleton, UK-based vendor comes just six months ahead of forthcoming European Union rules on spam that adds to mounting legislative pressures that are building in the US. The EU directive is to be implemented by October 31 into UK legislation by way of The Privacy and Electronic Communication Regulations 2003. It will mean that the rules governing email distribution will be changed from opt-out to opt-in so that people must actively choose to receive a communication from a company.
Businesses that fail to comply could face enforcement proceedings from the Information Commissioner including but not limited to a fine of up to 5,000 pounds ($8,063) per offence. Personal liability may also attach to company directors, managers and/or secretaries for such offences, it is said.
In the US the so-called Can-Spam Act (Controlling the Assault of Non-Solicited Pornography and Marketing Act of 2003) would allow state attorneys general to sue spammers on behalf of the public. It would also allow the Federal Trade Commission to levy fines.
Source: Computerwire
