Résumé
This history of expressive rendering and modeling is largely one of imitation: we make paintings in the style of Monet, cartoons like Dr. Seuss, watercolors, mosaic tile images, and so on. In each case, problems in representing or simulating the medium require that we be ingenious or devious or both: line drawings have aliasing problems; watercolors require stable but fast fluid simulation, etc. A few recent papers suggest a new and very promising divergence from this path: a search for media that are intrinsically suited to the kinds of representations and interactions that are natural both for a human and a computer rather than those that are naturally occurring in the physical world. Iwill discuss this evolution, some of the potentials of such new media in both rendering and modeling, and some of their potential pitfalls as well.