Man-machine communication

In his book on man-machine communication [120], Geiser gives a good overview on the topic. The chapters comprise theoretical aspects of information processing and ergonomical requirements as well as technical aspects of the computer system. The basic model of a man-machine system, shown in figure 4.8 , is very similar to our model (compare with figure 1.1 ).

  
Figure 4.8 : Structure of a man-machine system (Taken and translated from [120])

Geiser has reviewed several different models of human-computer interaction. A very short description of them follows:

The layered model [246]:

Norman has introduced a simple conceptual model with seven different layers of user action: On top, the user formulates a goal. In order to execute it, three layers are needed (top-down): planning, action specification, and action execution. To perceive feedback, another three layers are involved (bottom-up): perception, interpretation, and valuation of the system's state.

The 3-level model [283]:

Rasmussen's model is a conceptual model of the user, comprising three hierarchical levels of behavior: skills, rules, and knowledge (SRK). The layers represent different levels of abstraction in the human (cognitive) processing system. A more detailed description of this model follows in the next paragraph.

The GOMS model [60]:

The GOMS model describes the user's task representation as a conceptual model in terms of Goals, Operations, Methods, and Selection rules. It is especially useful for the prediction of the user's behavior, but it has a major drawback: Abnormal behavior of the user is not considered.

The keystroke-level model (KLM) [59]:

This model only considers one aspect of man-machine communication: How long does it take for an experienced user to perform a standard task without any error? The model's benefit is the quantitative measurement of observable parameters in man-machine interactions. Drawbacks of the KLM are that cognitive aspects are difficult to include and that in many cases the measurements performed are not very accurate.

The theory of cognitive complexity [156]:

Based on the GOMS model, Kieras and Polson have developed a theory in order to allow a quantitative analysis of the complexity of the man-machine dialogue. The theory's purpose is to support the design process of dialogues with respect to the expense and the transfer of learning, the execution time, and the user friendliness of the system. The cognitive complexity is defined as the content, the structure, and the knowledge that the user needs in order to use a device. Additionally, two different knowledge representation schemes have been defined: One for the knowledge of the user concerning the task, the other one for the device itself. Unfortunately, the predicted results could not be proved in practice, and the model has been refined several times by other researchers.

Esprit Project 8579/MIAMI (Schomaker et al., '95)
Thu May 18 16:00:17 MET DST 1995