applied is to use our scenarios to structure interviews with subject-matter experts, either singly or in a group. These subjects have been asked to comment on the likelihood of certain events occurring given alternative future system designs, to predict the effects of such events on their operations, and to discuss ways in which they would avert the occurrence of such events or compensate for them if they did occur. In applying this method, an effort is made to utilize subjects from different specialties, in order to elicit their different perspectives. Several scenarios have been explored (and refined) using this approach.
Conceptual walkthroughs using future "incident reports." Scenarios have also been used as the basis for "incident reports" in which a future incident is predicted, and is presented in the form of a formal report investigating that incident. These incident reports, with supporting documentation, are presented to participants (we have used groups of air traffic controllers, dispatchers and pilots) to consider as if they had actually occurred in some future system. The technique is used to structure a conceptual walkthrough by the participants, eliciting the ways in which the incidents might have been avoided, and how the system might be insulated against such occurrences or their effects. We have used this method to study how cooperative problem-solving in a hypothetical system can be facilitated, and how roles, responsibilities, procedures, policies and technologies must be designed to enhance performance and to make the system as error-tolerant as possible.
"Role-playing" conceptual walkthroughs. Another scenario was constructed to permit the observation of cooperative problem-solving more directly. Using this method, subjects are given the background and context of a scenario and the rules under which the system is operating. They are presented with the onset of an event and are then asked to "play out" the scenario as it occurs. The role playing is supported by a gaming board that represents the aircraft in particular ATC sectors. The participants can manipulate the gaming board to play out how the situation could evolve given different contingencies, actions, and interventions. Multiple participants debate among themselves different strategies for handling the situation. The methods by which they jointly resolve the problem are the data of interest. We believe this approach has considerable potential as a second method to conduct a conceptual walkthrough, eliciting additional insights concerning how various human and machine elements would interact in some future system.
We have applied all three of the methods described above, working with a total of 40 controllers, dispatchers, pilots and traffic managers. A sample of the resultant conclusions is presented below.
Changing Roles and Information Requirements. Changing roles by re-distributing authority (locus of control) has strong implications for the kinds of information and information displays needed to support these new roles. New ATM concepts change the roles of many of the people involved in the system. Under some current proposals under consideration, dispatchers will have more flexibility in route planning; flight crews will play a greater role in ensuring separation; and controllers will act more as monitors, making new kinds of decisions about when to intervene. If the changes created by these shifts in the locus of control, and in decisions concerning whether and when to intervene, are not accompanied by a corresponding shift in access to information, problems can arise.
Airlines and Flight Planning. One of the major problems with the current system is that the ATM system often has no access to information about the impact of its decisions on airline business concerns. As a result of this separation of authority and information, many of the decisions made by traffic managers and controllers are based solely on considerations regarding traffic flows and separation. Even when two solutions to a traffic problem are equally acceptable in terms of safety and traffic flow management, FAA staff generally do not have the information necessary to select a solution that is preferable to an airline in terms of its business concerns.
As a response to such problems, the FAA has been shifting the locus of control to the airlines where