Jean-Michel Hoc

Expert Software Design Strategies

Early studies on programming have neglected design strategies actually implemented by expert programmers. Recent studies observing designers in real(istic) situations show these strategies to be deviating from the top-down and breadth-first prescriptive model, and leading to an opportunistically organized design activity. The main components of these strategies are presented here. Consequences are drawn from the results for the specification of design support tools, as well as for programmers training.

Language Semantics, Mental Models and Analogy

Abstract The semantics of a number of programming languages is related to the operation of a computer device. Learning a programming language is considered here from the point of view of learning the operating rules of the processing device that underlies the language, as a complement to the learning of new notations, or a new means of expression to be compared to natural language. This acquisition leads beginners to elaborate a new representation and processing system (RPS) by analogy with other systems that are associated to well-known devices. During acquisition, beginners not only learn new basic operations but also the constraints of these operations upon program structures. Learning therefore concerns a basic problem space as well as abstract problem spaces within which planning takes place. The links between this approach to learning to program and a number of related works on learning to use software are underlined. Implications of these research findings in the programmer training are drawn.

Strategies in controlling a continuous process with long response latencies: needs for computer support to diagnosis

An empirical analysis of blast furnace conductors strategies in a simulation of the process is presented and its implications for the design of intelligent computer support is discussed. The rationale for the choice of this situation is the need for cognitive analysis of process control situations which are far from discrete state transformation situations for which information processing psychological models or artificial systems have been designed. The simulation method is justified by the results of the previous steps of the study (behavioral observations in the control room and interviews on tool use and knowledge representation). The strategies are described in terms of representations used and processing performed, their efficiency is evaluated, and correlations between strategic features and efficiency are examined. A number of hypotheses are put forward on types of computer support best suited to satisfying the conditions of implementation of the most efficient strategic features. The computer is seen as an instrument, where it operates as a colleague, rather than a prosthesis capable of replacing the human.

Some dimensions of a cognitive typology of process control situations

In the area of process control, there is a need to elaborate the dimensions of a cognitive typology, which can be of relevance to explain research results and define their domains of validity. To do so, it is not enough to describe environments, but rather the results of interactions between environments and operators, termed situations. This paper discusses the conditions for achieving a cognitive typology, presents some of its possible dimensions and shows its usefulness. Six main dimensions are examined: (1) Control scope (causal, temporal, and spatial). There is a clear relationship between this dimension and expertise, through the use of schematic representations. (2) Control directness (effects of interventions on the crucial variables of the process). When control is indirect, interventions can initiate complex causal chains. This reinforces the need for planning and supervision of the implementations of these chains. (3) Process information accessibility. When information on the crucial variables ...

Cognitive Control Levels in Air Traffic Radar Controller Activity

Classically, cognitive control is conceived as a set of levels activated according to certain psychological or situational conditions. However, studies on process supervision and a recent theoretical model suggest the possibility of a parallel activation on different control levels. We have analyzed the en route radar controller activity of conflict detection and resolution. Variables included aircraft position drawings during simulation freezing periods and temporal relations between activity steps and air traffic temporal dynamics. Data show evidence for a distinction between drawings indicating anticipatory mental representations and an accurate synchronization of cognition with the air traffic dynamics. Inferences are drawn for the design of a new support tool capable of implementing controllers abstract action plans.

An ergonomic approach to knowledge representation

Abstract The mutual contribution of an ergonomic (interview and simulation) and a technical (process supervision theory) approach to knowledge representation is examined. Initial validation of a causal model for diagnosis in the supervision of a continuous process with long response latencies (a blast furnace) is described. This validation shows that the model accounts for a high proportion of hypotheses elaborated during diagnosis/prognosis and that only features of the model which are directly linked to intervention are used in reasoning. The model is designed to specify an operational language for operator-computer and operator-operator communication during diagnosis. Discussion centres on the relevance of this model in accomplishing this goal, and ways of improving the availability of schematic representations of either the state or the evolution of the process, in planning.

ANALYSIS OF COGNITIVE ACTIVITIES IN PROCESS CONTROL FOR THE DESIGN OF COMPUTER AIDS: An example: the control of a blast furnace

This paper presents the preliminary findings of a research program designed to model diagnosis activity in order to propose a computerized support system for a continuous process characterized by long response latencies: the blast furnace. Part one examines process control from a cognitive point of view. This is followed by methodological considerations. The major features of the control activity are then defined in terms of its main components: supervision, diagnosis, and intervention. Conclusions as regards computerized support system design are put forward.

Cognitive Cooperation in Dynamic Situations: Lessons Drawn from Fighter Aircraft Piloting and Air-Traffic Control:

In many dynamic situations, which are highly goal directed and temporally constrained, most of the cooperation difficulties (e g., problems between humans and automation) can be reasonably addressed by adopting a cognitive viewpoint. Cooperation is taken as the management of interference between individual activities to facilitate the team members subtasks and the teams common task when there is one. Thus, planning models are of great relevance in identifying cooperation difficulties and supporting cooperative activities. Studies on individual activity in dynamic situations have stressed the importance of the elaboration and maintenance of a current representation of the situation, (an extension of situation awareness). The concept of Cofor (common frame of reference) generalises this idea to teams. Two series of results drawn from rapid process control situations where there is insufficient time to converse at length (air traffic control and double seat fighter aircraft) show that Cofor activities are first invested into the recall and confirmation of plan, task allocation, and decisions at key points before action completion The investment in prevention reduces the risk of fatal interference, or missed action