Although it seems like the spread of spam-unwanted junk e-mails sent to millions of people each day - is a recent problem, spam has been around as long as the internet has. In fact, the first document case of spam occured in 1978, when a computer company sent out 400 e-mails via the Arpanet, the precursor to the modem Internet. Now, spam e-mails account for more than two-thirds of all the e-mails sent over the Internet and for some unlucky users, spam makes up 80 percent of the messages they receive. And despite technological innovations such as spam filters and even new legislation designed to combat spam, the problem will not go away easily.
Most spam filters rely on a fairly primitive ''finger printing'' system. In this system, a program analyses several typical spam messages and identifies common features in them. Any arriving e-mails that match these features deleted. But the fingerprinting defence proves quite easy for spammers to defeat. To confuse the spam messages change ''fingerprint'' and thus allow the spam to escape detection. And the programmers modify the fingerprint software to look for random strings of letters, spammers respond by including non-random content, such as sport scores or stock prices, which again defeats the system.
A second possible solution takes advantage of computer's limited learning abilities. So called ''smart filters'' use complex algorithms, which allow them to recognize new versions of spam messages. These filters may be initially fooled by random characters or bogus content, but they soon learn to identify these features. Unfortunately, spammers have learned how to avoid these smart filters as well.