When I published TSPLIB in 1995, I expected
that at least solving the large problem instances to proven optimality
would pose a challenge for the years to come.
However, due to enormous algorithmic progress all problems
are now solved to optimality!!