The design of complex multimedia documents presents new challenges to authoring systems, because spatial and tem- poral features should be visualized and made accessible in an intuitive and direct-manipulative way. In this study, multimedia presentations are considered as hierarchical compositions of time objects that define serial or parallel synchronization of the inserted media objects. Media com- position hierarchies support automatic temporal layout mechanisms. They are integrated into an object-oriented application framework called MET++. It provides direct manipulation interaction on temporal structures in a compo- sition view that displays temporal and compositional rela- tions, allows time transformations on composites, supports the definition of time-dynamic behavior of media objects through time functions, and has a flexible graphical layout in order to visualize context information. Keywords: multimedia composition, audio, 3D graphics, time synchronization, animation, user interface interaction, object-oriented application framework, C++, MET++.