And how is the technology of MegaTextures working? I can think about it as dividing the mesh in several pieces (the visible ones), loading the texture fragment for the corresponding meshes and rendering (paging). Might not be that simple, or does it? Also, if U-V coordinates are values from Single(0.0) to Single(1.0), isn't it a problem to deal with that huge resolution (128000x128000 pixels) to map the meshes? There should be a limit regarding to floating point precission. Or is it possible to feed the U-V with practical values larger that 1.0? :?