The thrust of this research is the development of picture-coding techniques that are an order of magnitude more compact than existing codes for certain types of materials, and are also more easily manipulated. For example, we are attempting to store an entire year of a serial television program on a two-hour digital videodisc (DVD). In addition, since interactive video has suffered from a lack of "re-usability" (images seen once become boring on round two), this coding technique can make a video game where the same footage never reappears. The work is beginning with some long-term characterizations of television programs.