In '96 #3 's final (also HW5), some people had lots of constraints that
were circular and couldn't do anything with them.  Usually, this is messy
to figure out and it's mostly trial and error (cuz your merege diagram
would only work if some states are the mergeable).  However, you could
cross most of them out and only get state A and D.  This is done by,
looking at both constraints in a aquare.  For example, A=G if A=G and B=G,
which means A=B=G, but A != B if you look in the implication table, A=B is
crossed out, so you can cross out A=G.



In '95 #4b, there are don't cares in all of column 11, which is bad.
Also, C in row A and E in row C can be don't cares (because of fundamental
mode and no other state uses it as a transition state).



In '94 group 1, #3c, T1=(X Y2) XOR Y1 is a kind of clever solution to
part of the problem.