Switching from a traditional rasterization approach to something voxel-based just sounds horribly work-intensive. So many of the development tools would need to be rewritten... and I can't imagine how content creation would work.
Although I guess you could create content in a traditional manner and then run the high-poly models through some sort of converter... just sounds kinda nuts, though!
Good luck with all that. I'd love to see a demo... but I'm not holding my breath.
This comment was edited on Mar 12, 16:27.