Definitions and Objectives

Listed below are characteristics of what we are terming agents followed by an example of those characteristics as found in our system:

• Delegation -- e.g., the ability to receive a task to be performed without the user's having to state all the details
• Data-directed Execution -- e.g., the ability to monitor local or remote events, such as database updates, OS, or network activities, determining for itself the appropriate time to execute.
• Communication -- e.g., the ability to enlist other agents (including people) in order to accomplish a task.
• Reasoning -- e.g., the ability to prove whether its invocation condition is true, and to determine what are its arguments.
• Planning -- e.g., the ability to determine which agent capabilities can be combined in order to achieve a goal.

Our initial prototype includes agents that exhibit aspects of all the above capabilities, except planning (but see [[7]]). Our goal is to develop an open agent architecture for networked desktop and handheld machines. The system we are building should support distributed execution of a user's requests, interoperability of multiple application subsystems, addition of new agents, and incorporation of existing applications. Finally, it should be transparent; users should not need to know where their requests are being executed, nor how.