Generating an Abstract User Interface from a Discourse Model Inspired by Human Communication

Programming user interfaces is hard, error-prone and expensive, but recent advances allow generating them from interaction design models. We present an approach for modeling interaction design that is inspired by human communication. Our interaction design models are discourse models, more precisely models of dialogues. They are based on theories of human communication and should, therefore, be more understandable to humans than programs implementing user interfaces. The main ingredients of our models are communicative acts (Speech Act Theory), which are connected as adjacency pairs (Conversation Analysis) and via RST relations (Rhetorical Structure Theory). While RST provides useful means for modeling discourse in the sense of monologue, a dialogue results from connecting monologues via adjacency pairs. This paper presents a new metamodel that integrates these approaches. Based on it, we show how abstract user interfaces can be generated from such discourse models. In a nutshell, we generate finite-state machinery and employ rules devised by us to map parts of a discourse model to abstract widgets.