We propose a “Zooming Cross-Media” concept that uses zooming to achieve both changes in the level of detail and transitions between media, for contents containing varied media. Examples are text, images, video, and sound. As part of the concept, we propose a zooming description language (ZDL) based on XML. Unlike existing zooming interfaces, ZDL codes the zooming operation and behavior on the content side. Because ZDL adopts XML coding, we can locate “zooming” as the third interface in the Web document environment after “scrolling” and “anchor clicking.” The zooming operation and behavior is independently coded from the content structure in ZDL. With ZDL, it is possible to (1) control the zooming of each “zoom object” making up the contents, (2) control the degree of zooming by introducing a “zoom rate” parameter, and (3) relate objects mutually and specify zooming propagation between related objects.