Integrating video into an application framework

Abstract

Object-oriented application frameworks help software engineers develop sophisticated interactive applications quickly and reduce maintenance costs significantly. An application framework provides, among others, predefined visual classes for text or graphics display and for user interaction elements, as well as integrated control of these elements. Designed before motion video display was technically feasible on desktop computers, these frameworks’ architectures are built around the assumption of relatively slowly changing bitmap images to be displayed on a computer screen. Motion video screen update rates were not anticipated. We extended a typical application framework in a multitasking environment to support motion video display in full generality. The internal architecture of the framework was changed to remove the subtle obstacles for high screen update rates and a data type ‘video’ was designed and integrated seamlessly with the existing visual classes. Our video objects display motion video generated by autonomous hardware or software processes. They may appear in any shape or number mixed with other visual objects in scrollable views, as building blocks in graphics editors, as characters in text editors, as items in list or pop-up menus, etc. Video objects support the full visual class protocol for client-transparent double buffering, cut/copy/paste/undo operations, and output to or input from files. In this paper, we describe the features of these video objects and our changes to the architecture of the application framework necessary to support the new paradigms of motion video.