Recently, science museums have begun to review their educational purposes and redesign their pedagogies. At the most basic level, this entails accounting for the performance of individual exhibits, and indeed, in some cases, research indicates shortcomings in exhibit design: While often successful in prompting visitors to carry out intended actions, exhibits do not necessarily promote the intended interpretations of these actions among visitors. Here, the notion of "praxeology" from didactics research is suggested as a model to remedy this shortcoming. The suggested role of praxeology is twofold: as a means to operationalize the link between exhibit features and visitor activities; and as a template to transform scientists' practices in the research context into visitors' activities in the exhibit context. The resulting model of science exhibit engineering is presented and exemplified, and its implications for science exhibit design are discussed at three levels: the design product, the design process, and the design methodology.