Résumés
Résumé
Le processus de conception d'une situation de travail résulte de la coopération de plusieurs acteurs (opérateurs, concepteurs, ergonomes). À travers trois études, nous illustrons le fait que les résultats de l'analyse de l'activité sont essentiels pour orienter le processus de conception, mais aussi pour favoriser la coopération des acteurs. La première étude met l'accent sur l'intérêt d'une analyse de l'activité collective en vue de la définition des nouvelles salles de régulation du trafic urbain. La deuxième étude souligne l'intérêt pour les concepteurs de réaliser l'évaluation d'un système d'assistance en situation réaliste d'utilisation. La troisième étude montre le rôle moteur que joue l'analyse de l'activité aux différentes phases d'un processus de conception informatique.
Abstract
This article is based on three empirical studies carried out in different work environments: metropolitan traffic control, car driving and software design. The article illustrates how ergonomics can contribute to different stages of design processes.
Activity-centered design is described first. Designing truly supportive technologies requires an understanding of the user's actual activity. Therefore, it is necessary for ergonomists to collect data on natural work situations and to propose descriptions (sometime models) which are useful to cooperate with designers. Ergonomie studies may be carried out throughout the whole design process or they may be developed at specific stages of the process.
The first study emphasizes the significance of analyzing collective work in order to define future metropolitan traffic control rooms. The characteristics of the collective handling of disruptions in traffic control rooms orients the design of new tools favouring cooperation and communications. We briefly present one of the proposais we have put forward: a device enhancing the computer system that is able to support the individual monitoring of an incident, as well as the collaborative supervision of train movements.
The second study stresses the relevance of considering the evaluation of a support system in a realistic situation. The purpose was to evaluate how the drivers assimilated the "Autonomous Intelligent Cruise Control" System in their activities, and the actual support provided to them by the system. The methodology consisted in the observation of the drivers' behaviour during a highway trip of 900 km. The results of the activity analysis and the ergonomie approach allowed modifications of the designers' views on several points. In short, we have stressed the role of the driver throughout the design process, and demonstrated the limitations of the designers' views.
The last study shows the importance of activity analysis in the different phases of a software design process. This study is an important illustration of how participants cooperate and how this cooperation evolves at each step of the process. Cooperation between the participants is studied in relation to a particular programme (the definition of a design methodology for a software help package). Three phases are detalled: the definition phase of the functional requirements, based on the analysis of the activity, in order to favour innovating proposais; the specification phase, based on a modeling of the activity (in our case, the modeling of human help), aimed at a precise software model, complete and consistent with the user's needs; the design phase, aimed at the definition of technical solutions consistent with the functional requirements. This latter phase saw the design of a dialogue prototype from which an ergonomie experiment was developed.
In conclusion, we stress the need for collaboration between all the participants in the design process: users, ergonomists and designers who will need to confront and coordinate their points of view in seeking to transform the "work situation". Such cooperation has a strong influence on the behaviour of all the participants: the designer should broaden his or her approach, and should not adhere to a purely technical vision; the user, on the other hand, is not a mere executant, and should be involved in the different phases of the design; and the ergonomist cannot be concerned only with the end product, but has to be involved as early as possible in the design phase.