constraints. But as we pointed out at the start, if any information system is to be effective, it must integrate environment and cognitive constraints. How can we retain the value of both types of constraints, while still acknowledging the primacy of environment constraints?
Our answer to this question can be explained with reference to Fig. 2.6. The ecological approach to work analysis begins by analyzing physical reality (e.g., nuclear physics) and social reality (e.g., the intentions of a remote collaborator), whereas the cognitivist perspective begins by analyzing workers' cognitive characteristics (e.g., mental models, strategies, preferences). At the risk of being overly repetitive, we reemphasize that what is under consideration is which set of constraints should be given precedence, not which set should be the subject of work analysis--both are important. Having said this, the basic point of this section is that cognitive compatibility is of little use unless ecological compatibility has already been established. That is, making an interface compatible with a worker's mental model will not do much good if that model does not correspond with the way the environment actually functions. It is for this reason that environment constraints need to be given priority.
However, this does not mean that there is no room for cognitive constraints. On the contrary, attention to human characteristics is of the utmost importance and should also be given an important place in work analysis. Once it has been established that environment constraints have been respected, knowledge of human characteristics provides an important basis for making the remaining design decisions. In this way, work analysis can provide a way of making sure that physical and social reality are not ignored, while simultaneously ensuring that cognitive characteristics are not overlooked. The end result should be computer-based systems that will present workers with the information they need to develop accurate mental models in a form that is compatible with existing knowledge of human cognition. The framework that we present later in this book embodies this integrative approach, synthesizing environment and cognitive considerations while giving primacy to the environment.
In this section, we have presented four examples from the nuclear industry to show that, even in this patently correspondence-driven domain, researchers, designers, and regulators have not taken the primacy of environment constraints as strongly as they should. A number of important limitations arising from this oversight were discussed. We concluded that a work analysis framework for complex sociotechnical systems must integrate environment and cognitive constraints while giving primacy to the environment.
This book is primarily about analysis, not design. But as we pointed out in chapter 1, if analysis has any pragmatic value at all, it must be in the implications it has for design. The first part of this chapter showed that when computer-based information systems are designed without taking into account work demands, problems arise.