We have two different implementations of heap sort. Can we build a benchmark to prove which one is faster or better than the other? Benchmark against small datasets, large datasets, and empty datasets. It is OK to keep both if they have different strengths and weaknesses. If there is a clear winner, let's delete the other one.