From surface form to the structure of the interface — studies in human computer interaction at INRIA

Abstract

The central problem in man-machine interaction is the compatibility between two elements:

• (a) the operator, and his physical, perceptual and cognitive characteristics;

• (b) the machine, and its different aspects: dimensions and lay-out, information coding and information structure.

At a first level, the designer must endeavour to attain a certain compatibility between the physical characteristics of the system and the physiological and perceptual characteristics of the human. At this level, the designer is concerned with work place dimensions and general lay-out, information visibility, etc. This field — ergonomics — is now very well established. At a second level, the designer's task concerns the compatibility of the system with the elementary operations performed by the operator. These operations consist of acquired schemes, which can be sensorimotor, procedural and / or anticipatory. This rule-based behaviour must be matched by appropriate surface aspects of the machine. The designer has to choose, for each sub-task, the optimal way to encode information in order to facilitate the use of these schemes. To give an example, for a tracking task, different types of displays will be studied (pursuit, compensatory, predictive, analogical, pictorial, etc.). The studies of stereotypes (which are sensori-motor or cognitive routines) belong to this second level of compatibility. This is the field of human factors. Finally, at a third level, the designer has to take two fundamental human activities into account: information processing and mental representation. The relevant questions become: which information is processed? Which variables are elaborated by the operator? Which heuristics, which reasoning algorithms are used to reach the goal? What are the characteristics of the mental representation of the system? At this level, studies of the knowledge activities are fundamental. On the machine side, the designer is no longer concerned with information encoding, but with information structuring. This is what cognitive engineering is about. The global compatibility of a man-machine system can be achieved only if compatibility exists at each of these three levels.

As a matter of fact, the available body of knowledge is very unbalanced. We know a lot about ergonomics, quite a bit about human factors, and not much about cognitive engineering. The aim of the Ergonomic Psychology Project at INRIA is to contribute to the development of knowledge on the cognitive activities of the human elements of the systems: this knowledge can be used to design machines adapted a priori to their users' functioning logic. Our work is then clearly focused on the third level of compatibility described above. However, it is sometimes difficult to have a clearcut separation between the different levels, especially between the second and third one. In fact, surface form and deep structure interact in several ways. The aim of this text is to present some aspects of these interactions, which will be illustrated by examples from different studies we have conducted. We are concerned with all problems related to human-computer interaction, whatever the domain of application: process control, office work, data base interrogation, programming, etc.