Working on encoding Beck-Chevalley into the blob
of twelf I already have. Added some monic
manipulation lemmas and defined the arrow part
of the power object functor and direct image.
Hm. Might map_c follow from map_u?
Maybe not.
Garg, bumping up against performance issues
trying to prove the two-pullback lemma.
There must be a better way...
Car is fixed after all. Huh.