Contact us Heritage collections Image license terms
HOME ACL ACD ICF SUS DCS G&A STARLINK Literature
Further reading □ Foreword □ Contents □ Preface □ Prologue □ Acknowledgements □ Participants □ 1. Introduction □ 2. Control Structures □ 3. Syntactic Structures □ 4. Cognitive psychology and interaction □ 5. Visual Communication □ 6. Presentations □ 7. Working Groups □ 8. Group Reports □ 9. Postscript □ 10. Position papers □ 10.1 Anson □ 10.2 Baecker □ 10.3 Bo □ 10.4 van den Bos □ 10.5 Crestin □ 10.6 Dunn □ 10.7 Dzida □ 10.8 Eckert □ 10.9 Encarnacao □ 10.10 Engelman □ 10.11 Foley □ 10.12 Guedj □ 10.13 ten Hagen □ 10.14 Hopgood □ 10.15 Klint □ 10.16 Krammer □ 10.17 Moran □ 10.18 Mudur □ 10.19 Negroponte □ 10.20 Newell □ 10.21 Newman □ 10.22 Nievergelt □ 10.23 Ohsuga □ 10.24 Rosenthal □ 10.25 Sancha □ 10.26 Shaw □ 10.27 Tozzi □ 11. Bibliography
C&A INF CCD CISD Archives Contact us Heritage archives Image license terms

Search

   
⟶ ACD ⟶ Literature ⟶ Books ⟶ Methodology of Interaction
⟶ ACD ⟶ Literature ⟶ Books ⟶ Methodology of Interaction
ACL ACD C&A INF CCD CISD Archives
Further reading ⇓

□ Foreword □ Contents □ Preface □ Prologue □ Acknowledgements □ Participants □ 1. Introduction □ 2. Control Structures □ 3. Syntactic Structures □ 4. Cognitive psychology and interaction □ 5. Visual Communication □ 6. Presentations □ 7. Working Groups □ 8. Group Reports □ 9. Postscript
10. Position papers
□ 10.1 Anson □ 10.2 Baecker □ 10.3 Bo □ 10.4 van den Bos □ 10.5 Crestin □ 10.6 Dunn □ 10.7 Dzida □ 10.8 Eckert □ 10.9 Encarnacao □ 10.10 Engelman □ 10.11 Foley □ 10.12 Guedj □ 10.13 ten Hagen □ 10.14 Hopgood □ 10.15 Klint □ 10.16 Krammer □ 10.17 Moran □ 10.18 Mudur □ 10.19 Negroponte □ 10.20 Newell □ 10.21 Newman □ 10.22 Nievergelt □ 10.23 Ohsuga □ 10.24 Rosenthal □ 10.25 Sancha □ 10.26 Shaw □ 10.27 Tozzi □ 11. Bibliography

10.12 Remarks on Some Aspects of Man-Machine Interaction

Richard A. Guedj

Thomson-CSF, LCR

Corbeville B.P. 10, Orsay 91401 France

Although some users are satisfied with some interactive systems the designers of such systems are unable to predict the quality, nor to specify the data to collect and analyze. A pragmatic and short term goal is to set up guidelines intended to improve the quality of interactive systems. An ambitious goat is a better understanding of interaction ; this may come after the building of a satisfactory model of interaction and the shaping of convenient tools to describe dialogues.

Remarks and comments are made pointing at the asymmetry of the dialogue, the need of a conceptual framework, some characteristics of the command language. Consequences are derived. Emphasis is put on the study of modes of communication

INTRODUCTION

The era of man-machine symbiosis had been announced almost two decades ago. The design of effective and symbiotically suitable man-computer interfaces has been pursued ever since. Looking at the results so far, it does not seem that we fully understand yet how to design interactive systems with the confidence and certitude that the man-machine interface is adequate. With many existing interactive systems, it seems that the designer (or team of designers) has been unable to set up a precise objective or criteria to evaluate the design with respect to the user-perceived quality of the system.

When embarking on the design of an interactive system, the situation of the designer, with respect to the quality of interaction is in general the following:

  1. the specific objective to be pursued is not determined
  2. a model of man-machine interaction has not usually been created
  3. the logical consequences of the model, when there is a model, have not been determined.

Some data are necessary for an evaluation of the quality of interaction. However the designer is unable to specify with confidence what data are required, in what form, to what degree of precision, in what amount, and how the data are to be collected and analyzed.

On one hand the state of knowledge about interactive systems is very low. On the other hand, interactive systems do exist with users expressing a fair amount of satisfaction. What should be our next move?

I suggest that in this workshop we adopt both a pragmatic and ambitious point of view.

From the pragmatic point of view, I would suggest to set up guidelines intended to improve the quality of interaction (whether we fully understand it or not). The ambitious point of view has to do with a basic and long term goal, to get a better understanding of the whole process of man-machine interaction.

I have a suspicion that this understanding may be nearer once we have walked two steps: the building of a satisfactory model of interaction and the shaping of convenient tools to describe dialogues.

APPROACH

The most important partner in the man-computer dialogue being the man, it seems natural to look first for such issues where it is possible to single out factors which have a measurable effect on user's behavior. Such issues have been compiled by several authors, Kennedy (1975), Miller and Thomas (1977) for an extensive list. The influence of a particular factor such as the system's response time, has been the subject of more detailed studies in specific contexts and systems. Miller and Thomas (1977) provide a recent review of such studies. Those issues seem to belong to a category of issues which yield a list of properties of the system designed. The system should show good facilities for editing, debugging, documenting, filing, inter-user communications, prompting, reliability, recovery, etc... It is difficult to organize and weigh those properties with respect to the quality of interaction. I tend to consider the corresponding studies as providing a necessary and important background with a wealth of information. However, for the present time, it seems that it might be more fruitful to shift the emphasis towards more global aspects.

By the way of general remarks on three aspects of the man-machine interaction process, I would like to draw the attention on what intuitively seems basic to an understanding of the whole picture. The three aspects considered are the asymmetry of the dialogue, the conceptual framework, the command language.

THE ASYMMETRY OF THE DIALOGUE

It seems rather trivial to remark that the computer is not a human-being and that the relationship is not a symmetric one. However, this has a rather important consequence for the handling of the dialogue. A sentence expressed by a human being, carries a meaning, i.e., conveys and reveals the significant intent of the author. When expressing himself, a human being is not only communicating with others, he also makes his statements for him to know what is his significant intent. This is done to add a new meaning to his knowledge as well as to the one of the listener. In short, the human being gives meanings.

Now, we make the assumption that our long term goal of understanding the interaction process, is quite distinct from the goal of Artificial Intelligence (AI), even if we expect the design of man-machine interfaces to benefit and use some AI techniques. As Treu (1975) very aptly put it: The basic objective is not to cause the computer to simulate human conceptualization and the cognitive processes leading to communication. After all, the inside of a computer is still as different from the human mind as it was when Simon and Newell (1964) described it. Instead, in considering the man-computer partnership, and given the current states of knowledge about the functionings of man and the computer, the objective is simply to design the "computer-understood" command language to be more conducive to man's conceptualization and formulation?

The role of the computer is therefore to accept the expressions with their intended meaning. Then it must process the statements and tell in such a way that the human receiver (or group) finds the computer message sufficiently intelligible and important so that the dialogue may progress. In short the computer does not give new meanings. Therefore the main form of communication from the computer is more related to the concept of telling, see B.N. Lewis and Jenny Cook (1969) for a theory of telling.

To summarize, on one-hand, we need a computer-understood command language, facilitating man's conceptualization and formulation. On the other hand, the part of the dialogue from the computer program seems independent of the command language used. As a consequence, the idea of symmetry and reversibility, although usually enlightening and fertile,, does not seem to apply to the two sides of interaction, (i.e., input and output are not necessarily symmetrical).

IN SEARCH OF A CONCEPTUAL FRAMEWORK FOR INTERACTION

A conceptual framework for graphics systems design has been proposed at the IFIP, Seillac-I Workshop (1976). This framework seems to be a fruitful one and leading to guidelines on how graphics packages should be designed. I just want to point out that it is essential to have a conceptual framework in order to make progress in understanding. Moreover, the importance of an effective transfer to the user's mind of the conceptual framework that guided the design is strongly felt. As Bennett (1972) said with a very good insight in his review of user interfaces in interactive systems: In any event, the impact on system performance of the user's concept of the tool is too important to be left to chance.

Now, how do we go to build that conceptual framework? Among interactive systems users and designers the idea seems to emerge that this framework could be derived by a careful study and subsequent taxonomy of user's tasks. What are user's tasks?

Treu (1975) distinguishes at least five types of computer-aided tasks (editing, programming, retrieval, design, instruction). Miller and Thomas (1977) tend to see two classes of human activities in routine tasks or problem solving tasks. Kennedy (1974) considers the purely practical aspect of a user communicating to a machine either (a) to add to its corpus of knowledge or (b) to obtain advice or information from it.

It is likely that other classifications could be made. A difficulty seems to be to find non-overlapping categories which remain natural.

A suitable distinction also should be drawn between tasks and modes of communication. For instance. Miller and Thomas propose four modes of communications (called styles of dialogue) in terms of two independent characteristics:

  1. Whether the interchange is guided overall by the user of the system, and
  2. whether the user has to make a choice of his input from a set of alternatives presented by the system or else is able to provide a free response.

Obviously we are in need of a good theory of human communication in order to make progress. Will it be possible to establish useful general purpose classifications of user's tasks or activities? Will this lead to the identification of natural modes of communication? Let us suggest to adopt two criteria which were given by B. Lewis and Cook (1969) to apply to modes of communication:

THE COMMAND LANGUAGE

From the wealth of information already gathered on the use of command languages, we already have a long list of properties that should be exhibited. I will only single out some which seem essential.

  1. It must be a computer understood language.
  2. It must facilitate man's conceptualization and formulation.
  3. It should not be too complex in order to be learned easily.
  4. It should be rich enough to allow for a powerful expression.
  5. It must be flexible enough to allow the user to engage in different modes of communication.

To characterize all types of user's commands, Treu (1975) has suggested an interesting scheme using what he defined as action primitives.

Those actions primitives defined via some general variable-field syntactic format have four elements:

  1. Action verb
  2. Action qualifier(s)
  3. Object(s) of the action
  4. Object qualifier(s)

This scheme or any other one should be measured with respect to the five properties mentioned above.

CONCLUSION

The three aspects of man-machine interaction which have been considered here, namely, the asymmetry of the dialogue, the search for a conceptual framework, the design of the command language, seem to play an important part for an understanding of the interaction process. All point to a need for an identification and classification of the modes of communication the users would like to engage in, when dealing with computer-aided tasks. Those modes of communication will be as basic to interaction as instructions and data were to a Von Neumann machine.

REFERENCES

Bennett, J.L. (1972): The user interface is interactive systems. Annual Review of Information Science and Technology. Ed. by C. Cuadra. Vol. 7, pp. 159-196.

Kennedy, T.O.S: The design of interactive procedures for man-machine communication. International Journal of Man-Machine Studies (1974). 6, 309-334.

Lewis, B.N. and Cook J.A.: Toward a Theory of Telling. International Journal of Man-Machine Studies (1969). 1, 129-176.

Miller, L.A. and Thomas, J.C. Jr. ; Behavioral issues in the use of interactive systems. International Journal of Man-Machine Studies (1977). 9, 509-536.

Simon, H.A., Newell, A. (1964): Information processing in computer and man. American Scientist. 52, 281-2.

Treu, S.: Interactive Command Language Design Based on Required Mental Work. International Journal of Man-Machine Studies (1975). 7, 135-149.

⇑ Top of page
© Chilton Computing and UKRI Science and Technology Facilities Council webmaster@chilton-computing.org.uk
Our thanks to UKRI Science and Technology Facilities Council for hosting this site