Specifying And Prototyping Dynamic Human-Computer Interfaces For Stochastic Applications

Formal methods are increasingly being used to support the software engineering of complex systems. A number of limitations restrict the utility of these techniques for the design of human-computer interfaces. Firstly, formal notations frequently abstract away from the temporal properties that aaect usability. Secondly, speciications often fail to consider the stochastic, or probabilistic, behaviours that characterise human-computer interaction with distributed and concurrent applications. This paper proposes techniques to overcome these limitations. It is argued that temporal logic provides a means of representing sequential and concurrent properties of interaction. It is also demonstrated that logic can be used to explicitly represent assumptions about operator responses to high and low risk events. In order to support the validation of these assumptions we have implemented a tool which exploits Monte Carlo techniques to directly derive prototype simulations from temporal logic speciications of interactive systems.