Hi Ravi:

a useful special case for negation would be if
the strategy is a layered graph but the class graph
is general. Does your idea of doing the negation construction
layer by layer work in this case?

See you on Monday at 2.30.

-- Karl