Help us improve direct modeling: record yourself modeling this camera!

Great. Now how about doing that automatically behind the scenes whenever you use sub object select to move a face (or control point) on a closed polysurface, so users don’t have to rely on Rhino’s super flaky history?