[Pitch] Selection: Linear-Time Order-Statistics

[rayx (Huan Xiong)]

@vanvoorden

To generalize our operation we want to select the element at any index k such that the k elements that appear before the element at index k are sorted as less or equal and the elements that appear after the element at index k are sorted as greater or equal.

I also find the term index is quite confusing because it implied the result is an array, instead of the k-th value. For example, your heap based implemantion doesn't return such an array. Also, index is 0-based, but k in select algorithms is 1-based. You were aware of that (see the part I highlighed) but it was confusing when I read it.