One interesting observation of 'modern' sorting algorithms that ship with programming languages is that they're mostly an amalgamation of various classical sorting algorithms with tests and heuristics to dynamically switch between them depending on the case. Cases like: few elements, elements already sorted with some out of order elements appended, elements sorted in reverse, detecting worst case behavior of quicksort and changing to heap sort instead, exploiting exit conditions of prior recursion steps to skip checks, etc etc. These kinds of meticulous opportunistic optimizations give us a chance to get theoretically ideal performance in both the worst and best case scenarios, and I hope we can apply this same methodology for many more algorithms in the future. Databases query optimizers sorta do this, but they typically make such decisions ahead of time so are prone to getting stuck with a bad plan. (Obviously this is highly dependent on your engine of choice.)