Interface transparency occurs when the user's
efforts are entirely directed at a task, and not at the
interface that supports that task. Transparency is vital
in many situations, particularly where a great deal of information
must be quickly integrated, and multiple tasks coordinated;
but at present, many important classes of interfaces are
not highly transparent.
Designing a transparent interface means knowing
what the user needs, how much and how quickly the user can
process supplied information in a given situation, and how
to adapt the interface to this context. In our opinion,
mainstream design of realtime, reactive user interfaces
has been hindered by the lack of a systematic process for
their design and validation, and for integrating different
user sensory and control channels. Our solution to this
problem will therefore be the creation of a systematic knowledge
base about human abilities in the realtime context, beginning
with the study of perceptual, attentional and task
transparency.
This project pursues this goal with three
interdependent research threads:
1) Realtime Platform: design a local
network cluster which coordinates sensor and display servers
with data and user model servers at the high communication
bandwidths, and supports a variety of refresh rates and
server operating systems.
2) Interplay between conscious and nonconscious
processing: Determine the characteristics of both types
of process in the visual, auditory and haptic domains by
extending existing psychophysical experimentation techniques.
3) Transparent adaptive user modeling and
interface prototyping: Iteratively design and evaluate
interfaces that adapt conscious and nonconscious user feedback,
including that perceived at "rapid" rates, through sensing
and interpretation of user state.
