In an interactive communication system based on MPEG-4, Command descriptors along with Command Route nodes or Server Routes in the scene description can be used to support application-specific interactivity. Content selection can be supported by specifying the presentation in command parameters, with the command ID indicating that the command is a content selection command. An initial scene can be created with several images and with text that describes a presentation associated with an image. Associated with each image and the corresponding text is a content selection descriptor. When a user clicks on an image, the client transmits the command containing the selected presentation and the server starts a new presentation. The technique can be used in any application context, as generally as HTTP and CGI can be used to implement any server-based application functionality.