add case for underspecified stack state (rule RHS has more elements than the stack state it is supposed to apply at)

introduce intentional non-monotonicity in order to verify that our normalization scheme for state numbers works correctly (we should also do the same for rule numbers)

print out extracted data; all seems good; we'll work on converting it into a bona-fide table and then convert that into a graph per our previous program

work on interface with bison in order to get IELR(1)-generated tables for testing the conversion algorithm