Resumen
The ability to develop future multimodal systems depends on knowledge of the natural integration patterns that typify people's combined use of different input modes. Given the complex nature of users' multimodal interaction, cognitive science will play an essential role in guiding the design of robust multimodal systems. In this respect, a multidisciplinary perspective will be more central to successful system design than it has been in traditional domains previously tackled by computer science. Specific design challenges will include developing multimodal architectures that can handle the time-critical nature of parallel interdependent input signals, as well as ones optimized for error avoidance and robustness. Ten myths regarding multimodal interaction have been identified and discussed from the viewpoint of contrary empirical evidence. In separating myth from reality, the goal has been to reveal the nature of multimodal interaction more clearly, which in turn provides a better foundation for guiding the design of future multimodal systems.