Ellen M. Voorhees
The advent of large wide-area networks connecting many diverse repositories of data creates the challenge of finding particular data of interest in an easy and timely manner. This paper describes the initial prototype of a system that addresses this problem by using a set of autonomous, customizable software agents. Central to the system design are scripts -- arbitrary, parameterized program segments written in a language that contains primitives for interacting with the data environment and tagged with keywords describing their purposes -- that define the functionality of an agent. Agents can query the keywords associated with other agents’ scripts and can import scripts from one another, thereby providing a mechanism for agents to learn from one another.