Multi-agent systems are rapidly becoming one of the most relevant success stories in the area of Artificial Intelligence, in that agent-oriented engineering is foreseen as the next dominant paradigm for the design and development of complex software systems. This introduces in the DAI field new issues concerning methodologies and enabling technologies, and calls for new metaphors and abstractions specifically supporting the engineering of MAS. In particular, the crucial point in MAS engineering is building agent societies. So, coordination models and languages are required, which allow the agent interaction space to be managed so as to rule social behaviours and accomplish social tasks. In this context, we argue that an effective coordination technology is needed, which not only implements a coordination model, but also supports its metaphors at the development system level, and enables MAS designers and developers to work at any time at the abstraction level they need.

In this perspective, this paper presents the LuCe coordination system for the design and development of Web-based MAS. Based on the full integration of Java and Prolog technologies, LuCe implements a coordination model based on the notion of logic tuple centre, a coordination medium which is also exploited as the core of the development system. The power of the LuCe coordination technology is first discussed in general, then shown in the context of a simple MAS: a TicTacToe game among intelligent software agents and human players on the Web.