Why stick to just prose or just music or just newspaper or just video? Why not create new models for information that combine elements of them all?