General Terms

Maes covers the topic of artificial life in the field of entertainment. Artificial life and autonomous agents are presented briefly in the introduction, and Maes explains why the application area of entertainment is regarded as important. Half of the paper surveys current agent research for animated movies and real-time animation systems. Common features of agents active in these systems are identified. Maes presents the standard research issues as well as more novel questions to be addressed.

The paper's second half presents a detailed description of the ALIVE project, which aims at building entertainment agents. The virtual worlds of ALIVE and the agents that inhabit them are described. Maes makes important observations of users' interactions with the system: the importance of intuitive user gestures; the need for an artificial guide; user tolerance during lags and occasional missed recognition when animal-like or humanlike agents are involved; and the importance of user visualization of the motivational and emotional state of the agent. Maes concludes with the observation that the particular interactions in which the user engages can be more important than fancy graphics.

The paper is well written, clear and concise. It includes several figures of the presented systems and a sufficient list of references.