Leveraging a better interface language to simplify adaptation

We describe an approach to building adaptive groupware systems. This approach encompasses a methodology that reduces the complexity of inferring user intent by identifying a domain-specific interface language that both supports the user's maintenance of common ground, and can be used to drive an adaptive component. Our approach can be framed as follows: 1) Users of same-time different-place collaborative systems must exchange certain types coordination specific information; 2) We can facilitate the exchange and management of this information by introducing special purpose interface components, which we call Coordinating Representations, that structure these communications; 3) Information that is collected through these interface components is particularly well suited to driving intent interencing procedures; and 4) Intent inference can be used to drive adaptive components that support the collaborative activity. We discuss empirical results from two experiments that validate this methodology.