Seems like internally asking "what would this look like if we could visualize it well?" is resulting in better visualization than the direct approach. -iore