Software agents are programs designed to perform tasks autonomously. Mail agents attempt to provide useful functions about electronic mail (E-mail) services, such as information filtering, gathering, and scheduling. The diffusion of the Internet has multiplied the amount of data and the number of information sources so the qualities that made E-mail so popular are now becoming a problem (e.g. the volume of junk or spam mail). Industrial as well as academic research has faced this problem in terms of automated filtering methods in order to distinguish, at the receiver-side, legitimate E-mail from spamming. We describe an alternative approach: our mail system is able to find, at the sender-side, "appropriate" destinations for a message by triggering a spidering process on (a portion of) the Web. This process performs distributed computation using mobile agents: by applying similarity-based reasoning on information extracted from the Web pages of potential addressees, the agents are able to compute a numeric value in [0, 1] which provides, for any address, a measure of the "interest" in receiving the E-mail.
展开▼
