Abstract Within a collaborative information environment, the system acts as a mediator between remotely located users. The user interface of such an information system has not only to provide support for the task to be performed, but also to reflect the ongoing discourse context. The MACIS Framework has been developed to support the design and implementation of collaborative, interactive information systems. Using a multi-agent approach, the whole information system, including its user interface, is decomposed in terms of various classes of cooperating agents. Since the users are also modeled as part of the multi-agent organization, interactions between all constituent agents are described in a uniform way. 1 Introduction Intelligent information systems are characterized as interactive systems which support various tasks using knowledge about the users and their task-related actions. To support collaboration between users imposes additional requirements on the underlying technical infrastructure. The MACIS Framework (Multiple Agents for Collaborative Information Systems) is designed to support the development of complex, collaborative information systems exhibiting intelligent, pro-active behavior. The system, including its user interface, is described in terms of various classes of cooperating agents. Since users are also modeled as part of the multi-agent society, collaboration between users is seamlessly incorporated as a special form of inter-agent communication. The user interface comprises dedicated user interface agents (UI agents), which serve as intelligent mediators between the users and the system. Hence, user input is no longer regarded as some method invocation, but as a communicative act expressing the current discourse goal of the user. Since users and agents are situated within a shared context, their interdependencies and interrelations are organized according to their roles within the information system.
[1]
Michael Wooldridge,et al.
Agent-based software engineering
,
1997,
IEE Proc. Softw. Eng..
[2]
Pattie Maes,et al.
Agents that reduce work and information overload
,
1994,
CACM.
[3]
Ben Shneiderman,et al.
Direct manipulation for comprehensible, predictable and controllable user interfaces
,
1997,
IUI '97.
[4]
C. Rich,et al.
Adding a collaborative agent to graphical user interfaces
,
1996,
UIST '96.
[5]
Joëlle Coutaz,et al.
Agent-Based Architecture Modelling for Interactive Systems
,
1996
.
[6]
Peter Sommerlad,et al.
Pattern-Oriented Software Architecture
,
1996
.
[7]
Nicholas R. Jennings,et al.
Intelligent agents: theory and practice
,
1995,
The Knowledge Engineering Review.