This paper explores the use of multi-actor virtual reality systems to
support communication and cooperation between groups of people. In par
ticular, it introduces a spatial model of interaction which aims to su
pport people in using their natural social communication skills when w
orking together. First, we briefly review some of the social science b
ackground which has informed the development of this model. Next, we d
efine the model's main concepts, i.e. aura, awareness, focus, nimbus,
adapters and boundaries. We also explore the issues of user embodiment
and spatial structure, the latter leading to the concept of Populated
Information Terrains. We then show how a few simplifying assumptions
results in the notion of awareness states and discuss how these link t
o the results of previous studies of cooperative work. Following this,
we present some early prototypes which realise our concepts within th
e DIVE distributed virtual reality system. In the first of these, a si
mple discrete version of the spatial model has been constructed based
on the notion of awareness states. The second, Q-PIT, demonstrates an
early Populated Information Terrain. Finally, we briefly reflect on th
e relationship between this work and other CSCW systems based on the m
etaphor of shared space.