I am not against using avatars, nor do I think they wont play a role in the future - in a few years they will dominate the web and other stuff, like intelligent assistents and so on. You are right taking this road. But there is still a lot of work in front of everybody involved in this business. My point is that the behavior is still rudimentary in its automatic generation - by the way we implemented the same stuff Keith Perlin has, on a Poser figure with Python. The figure makes all possible muscle movements in the face seperately and produces real good looking emotions. Real ones, not the handposed ones. This is currently used in research and will be available for the public sometime. We used Poser and everybody laughed at us when they saw the dork, but this has changed. you can see a real behavior example inclusive automatic breathing, blinks etc with a little bit of gravity added by Masa here: http://www.digitalmankind.com/virtual_fs.html Grammer