As we discuss below and elsewhere
[Turner, 1994][Turner, 1993], an event's importance depends to a very large
extent on the current problem-solving context; whether or not a response is
selected depends not only on the event's importance, but how reactive the
reasoner chooses to be in the current situation. Both of these factors can
depend on other agents or the overall problem-solving group. For example, the
failure of an AUV's light may have little importance for the AUV or its plans;
however, if it knows that a camera-bearing AUV is likely to ask it to
illuminate targets at some point during the mission, the event becomes more
important.
If an agent predicts that an event may have importance to the broader problem-solving system, it may need to inform others of the event, ask for information to help with the assessment, or even negotiate with other agents to settle on a mutually-agreeable importance estimate. The knowledge needed for this is similar to that for the other phases of event handling.