Stable topN?

[Andrei Alexandrescu via Digitalmars-d]

On 6/24/14, 11:50 AM, Xinok wrote:
>  From what I know of partitioning algorithms, they're either
> intrinsically stable or intrinsically unstable. Implementing an
> algorithm with the attribute you described would likely come with little
> benefit. I could be wrong, but this is my intuition.

I'd say start spinning some code and code will lead you to good insights 
regarding the relative costs of different stability guarantees. -- Andrei