The GML canvas: Aiming at Ease of Use, Compactness and Flexibility in a Graphical Toolkit

We present the GML canvas: a graphical toolkit aimed at supporting the development of post-WIMP interactions such as deformed, zoomable, semi-transparent, or bi-manual interactions. Our primary objective is to make the development of such interaction fast and easy; it is not to provide a library of post-WIMP interactions. The toolkit is used from an interpreted environment and generates its output with hardware rendering. We explain the rationale behind the design of the API and we present it in details. The API is split into a set of low-level atomic services and a set of high-level scripts. The atomic services offer the fundamental building blocks for interactive graphics and the scripts assemble these building blocks to serve as examples and to provide general interaction services. A number of examples illustrate how complex interactions have been built with a few lines of code.