When we learned about the Treesort algorithm, we saw that there was incompatibility between binary search trees and the keys we might like to sort: The keys might not be unique, but binary search trees operate under the assumption that keys are always unique (or, stated differently, if the same key is used more than once, it's assumed to be the same key). Propose how to solve this problem with Treesort; briefly explain what you would do to ensure that you obtained the correct result, even though binary search trees don't allow keys to be duplicated. Rememeber that we're not just sorting numeric keys, per se; we're sorting any one kind of object based on some characteristic of them, such as sorting strings by their length, employees by their years of education, or whatever, in which case our output would be the strings, the employees, and so on, not the characteristics by which we're sorting them.