A year ago, we published a special issue of The Knowledge Engineering Review entitled "Perspectives on Recent Intelligent Agents Research as Viewed through Two Conferences". Our intention was to provide a snapshot of current research on intelligent agents - an area which continues to grow apace - by looking at the work presented at the 1997 instantiations of two of the major events in the Agent calendar. These were the International Conference on Autonomous Agents (Agents '97) and the International Conference and Exhibition on the Practical Application of Intelligent Agents and Multi-Agent Technology (PAAM '97). The result was a group of six papers which together spanned the main areas covered by the conferences as well as identifying some of the challenges faced by the agents community as a whole. In particular, we had summaries by the organisers of both of the conferences, speculation on the future of robotics, surveys of expert assistants and electronic commerce, and a discussion of how agents could be used in workflow management.

Often, the problems encountered by applications and developments of software agents require further investigation and improvements of agents' social capacities, that is, of the properties enabling them to act adaptively in multiagent domains. This is so for at least three reasons. First, quite often software agents are designed for reasoning upon and interacting with other agents (both human and artificial), as is the case in client-server and personal assistants. Secondly, to accomplish their primary task, these systems ought to reason upon, predict and eventually coordinate with, or at least prevent obstacles from, other systems accomplishing the same or different tasks in a concurrent way (think of electronic commerce, meeting agents, etc.). Thirdly, many applications are actually distributed over a set of agents with their specific roles and tasks, which are expected to participate in a joint activity, aimed at achieving a given global outcome.

The rapid development of the field of agent-based systems offers a new and exciting paradigm for the development of sophisticated programs in dynamic and open environments, particularly in distributed domains such as web-based systems of various kinds, and electronic commerce. However, the speed of progress has been such that it has also brought with it a new set of problems. This paper reviews the current state of research into agent-based systems, considering reasons for the way the field has grown, and pointing at the way it might continue to progress. It pays particular attention to problems with defining the nature of agents, the technologies that have enabled the rapid progress to date, and ways in which work can be consolidated through the development of large-scale applications, and the integration with theoretical foundations.

This paper sets out, ambitiously, to present a brief reappraisal of software agents research. Evidently, software agent technology has promised much. However, some five years after the word ‘agent’ came into vogue in the popular computing press, it is perhaps time the efforts in this fledgling area are thoroughly evaluated with a view to refocusing future efforts. We do not pretend to have done this in this paper—but we hope we have sown the first seeds towards a thorough first five-year report on the software agents area. The paper contains some strong views not necessarily widely accepted by the agent community.The views expressed in this paper are those of the authors only, and do not necessarily reflect the views of our organisation.

How free are our software agents to take the best possible care of our interests? How free can we make them? In what sense and to what extent do currently proposed mechanisms and agent behaviors consider self-interest? What current research addresses these issues? What needs to be done? We address these issues in the context of electronic markets, such as consumer goods markets and (future, more fine-grained) markets for electric power or communication bandwidth.

When I began studying for my PhD, I knew exactly what field I wanted to work in. It was a subfield of Artificial Intelligence (AI) that was known as Distributed AI (DAI). In DAI, we studied how a number of (semi-)autonomous problem solving systems known as ‘agents’ could be made to cooperate and coordinate their activities to produce effective overall system behaviour (Bond and Gasser, 1988). At the time, not many people were working in distributed AI (only two books had been published on the subject). I found that when I talked about my research, I had to repeatedly explain what ‘agents’ were all about. It was depressing business. More than once, I wished I had gone into a more glamorous area of research - the expert systems people were getting all the glory.

The paper introduces a software architecture for the specification and verification of knowledge-based systems combining conceptual and formal techniques. Our focus is component-based specification enabling their reuse. We identify four elements of the specification of a knowledge-based system: a task definition, a problem-solving method, a domain model, and an adaptor. We present algebraic specifications and a variant of dynamic logic as formal means to specify and verify these different elements. As a consequence of our architecture we can decompose the overall specification and verification task of the knowledge-based systems into subtasks. We identify different subcomponents for specification and different proof obligations for verification. The use of the architecture in specification and verification improves understandability and reduces the effort for both activities. In addition, its decomposition and modularisation enables reuse of components and proofs. Therefore, a knowledge-based system can be built by combining and adapting different reusable components.

The continual growth of the Global Information Infrastructure (GII) prompts dramatic changes in both the types of information involved, and the kind of people who access it. In this paper, we briefly survey the most recent results with respect to the User Interface area, as emerged from the work presented at the fourth edition of the Conference on Advanced Visual Interfaces (AVI'98), focusing on interaction with large information spaces, multimodality, and usability issues.