Representing Appearance Information in a World of Interchangeable Documents

Ethan V. Munson

Abstract

Computer science currently lacks a widely applicable definition for the concept of a medium. Most existing definitions are implicit and treat data types, such as video streams, text, and digital audio, as examples of media. This viewpoint does a poor job of discriminating between many pairs of media, such as 2D static graphics and 2D animation, which share the same set of data types. Furthermore, this viewpoint is not suitable as a conceptual basis for software services that can be configured to serve many different media.

This paper presents presents a new, operational model of media. This model asserts that a medium is a triple M = (T, D, O), where T is a set of primitive data types (such as video streams or spline paths), D is a set of dimensions in which layout is performed, and O is a set of formatting operations with typed parameters. The model is sufficiently rich to discriminate between all commonly used media and can also be used to make statements about the extent of differences between media.

The model also has a practical use. It is the conceptual basis for Proteus, a software tool that manages the appearance and layout of multimedia documents. This tool uses a description of a medium to adapt its generic formatting services to that medium's needs.

Full text in PostScript.


This paper will appear in the Proceedings of the Third International Workshop on Principles of Document Processing, Palo Alto, CA, September 1996.