In this paper we describe how a set of autonomous computational
agents cooperate in providing efficient management
of information as well as control for the design
of a general and complete hypertext environment. We
discuss two populations of agents. The first population
is defined as support for the basic resources involved in
the hypertext; their duties are explained, proving how
appropriate distribution of task and knowledge enhance
the effectiveness of our model, in terms of control and
data management. The second category of agents is then
introduced; we explain how such category implements information
retrieval facilities and improves key issues of
hypertext systems, such as versioning management. The
computational model chosen to define the agent entities
is Agha's model; the problem of describing implementation effort is faced using a simple actor-level descriptive language
