I am a developer at the Jet Propulsion Laboratory in Pasadena working on planning software for the Mars rovers. Our product, Maestro, has an image browsing component which lets rover drivers and scientists examine pictures taken by the rovers, either as individual images or stitched mosaics composed of potentially hundreds of images. We use GEF to display our images and allow users to place target markers on points of interest therein. We wanted to give the users the ability to smoothly zoom in and out of a mosaic, allowing them to "fly" around the image in a Google Earth-style animated fashion. We quickly discovered, however, that the SWTGraphics which Draw2D uses by default was incapable of rendering scaled images fast enough to give the glassy-smooth transitions we were looking for, especially on the Mac and Linux platforms, which we must support. I decided to try implementing Draw2D's Graphics interface using OpenGL drawing routines, and swap this into the GEF/Draw2D framework, so that we could have our smooth animations of scaled images without losing the power of GEF for the target interaction layer. This talk discusses the approach I took to replacing Draw2D's graphics engine with an OpenGL core (without modifying the Draw2D plugin itself), as well as the several snags I encountered along the way to a successful implementation.