Discourse management in intelligent, multimodal user interfaces

Recent advances in software and hardware are allowing researchers to implement user interfaces that are able to communicate with the user in multiple modes, such as text, video and audio. Many multimedia interfaces are knowledge based as well. With the richness of presentation modes available, the discourse manager must make many complex decisions to choose the appropriate mode or modes to be used in each particular instance. The authors discuss the requirements of such a discourse manager, describe the knowledge sources needed to fulfil the requirements, and offer a sample blackboard architecture to illustrate its implementation.<
>