Hello, my name is Nick Cohea, and I am working on the indoor-outdoor unified rendering pipeline with Max Haagen-Loy. We have a repository (I am the user __queerty, the one with all the commits right now...) and a corresponding wiki.
We don't care about anything but culling triangles intelligently (lighting is a problem we want to treat for future work, probably as a deferred renderer. This is a forward renderer at the moment). Right now, the code uses the "strongly preferred" way of rendering, i.e., no immediate mode, display lists or anything like that, and we also use shader programs. I am going to be working in the graphics lab to use even more of the new OpenGL functionality, like occlusion queries and so on.
Our attitude toward the project is that the renderer shouldn't know much of anything in advance about the scene. We just take a scene, kill triangles, and send the ones that survived to the GPU. We will use every dirty trick to do this and do it well.
