Agent-based computing is a promising approach for developing applications in complex domains. However, despite the great deal of research in the area, a number of challenges still need to be faced to make agent-based computing a widely accepted paradigm in software engineering practice, and to turn agent-oriented software abstractions into practical tools for facing the complexity of modern application areas. In this paper, after a short introduction to the key concepts of agent-based computing and to the state of the art in the area, we try to identify a few key open research directions.