JOURNAL OF MULTI-CRITERIA DECISION ANALYSIS, VOL 5,
75
(1996)
Book Review
The world of Design processes and guidance (Chap. 7 )
is initiated with the classical spiral model from Hubka and is then oriented towards the mapping of the design territory. For example, we find there A map of the Cognitive systems engineering (CSE) offers a new knowledge representation to consider f o r information multidisciplinary framework for assessing needs and system design (p. 183) followed (in Chap. 8) by the designing usable, efficient systems (italics denote excerpts evaluation of design concepts and products, where the from the text). The dominant context is human- envelopes of constraints mentioned above are used as computer interaction and task performance in organiza- cognitive maps. The set is completed with the design of a librarian tions, while the disciplines providing the modelling concepts are engineering, psychology, cognitive science system and some references to health care institutions. Finally, the last part is devoted to the various visual and computer science. In CSE the main avenue to analyse the behaviour of displays and their contributions to cognitive systems systems is by abstraction to a functional level, which is design. As a whole, the 350 pages of the book make heavy, contrasted with structural decomposition where inputoutput models and causeeffect chains are often difficult reading. The book obviously offers new avenues of thinking and is definitely oriented towards considered in isolation. This leads, among other concepts, to a set of designs models, and, say, inquiring systems (a Churman defining behaviour shaping constraints at progressively expression) and not towards naive recipes or tools. However, almost all topics we can associate with any narrow envelopes. For the individual actor, for example, we find: Actual human activity system (in Checklands terminology) are coped with-metaphors, risk, propagation of conwork environment>Work domain analysis>Activity analysis> -in work domain terms> -in decision-mak- straints in an experimental design and a hundred ing terms1 -in terms of mental strategies that can be others-and although all of them are written under used>Cognitive resources analysis (actors competency, clever thinking (see the relevant Chapter 6 on human error), globally the mental burden is a problem for the criteria, values). For the activity analysis of an organization, such reader. Everyone reckons that human-machine interaction is embedding envelopes are (Chap. 3): Goals and constraints3Priority measures>Flows of values and mate- of very large scope and should, like the authors do, integrate many aspects, but not many people are rials>General functions and activities>Physical activities in work, vertically connected by the why, ready to swallow so many paradigms in such a dense flow. what, how questions. Thus a positive contribution is to open the mind to In decision-making terms, we find e.g. a decision ladder consisting of a sequence of basic information other mental designs than the common stuff so many processes in a decision task where the arrangement times reprinted, but the excess of scope and the difficulty implies: activation-observation--identification-options- of finding and following the authors path calls for a cerebrally muscled reader. goals-evaluation-prediction-choice of task-planningexecution. Control is presented (in Chap. 4) as a set of CHR.DE BRWN interactions between the levels of cognitive control Universitk de LiPge modes (i.e. knowledge-based, rule-based, skill-based) LiPge, Belgium and associated mental functions. COGNITIVE SYSTEMS ENGINEERING, Rasmussen, J., Pejtersen, A. M. and Goodstein., L. P., New York: Wiley, 1994.