During my Ph.D. thesis, I developed this system in order to be able to present my work (about image similarity and compression) in a more fanciful way.
The basic idea was to take standard PDF files (generated e.g. with Powerpoint, OpenOffice Impress or Latex-Beamer), and render them into an OpenGL environment, allowing detailed zoom into slides (e.g., into diagrams) and displaying e.g. 3D simulations and 3D statistics during the talk without having to switch to another window.
The system is currently able to display several sets of slides, physic-engine based spring embeddding animations, texture synthesis simulations, online image segmentation and fractal decompression, and can also be coupled with a web camera, in order to be able to acquire the images the algorithms should be applied to directly from the audience. (It was quite a lot of fun to do that in my final Ph.D. talk)