[nift]

Interesting sorting algorithm I haven't encountered before(even though it seems to be from 1998) m, but one that actually makes logical sense.

However, my first thoughts it does seem (from its concept) not so easy to implement(?). Additionally I would have concerns with regards to how much an overhead this calculation adds compared to just a “simple” comparison.

Maybe the calculation is worth it if the comparison is costly enough? My guess would at least be that we would need fewer comparisons in Flashsort as we should have a higher chance of “knowing” where things should go.

The Wikipedia article shares no plots/data (guess I should dig deeper for that), but would be interesting to see how well it fares against more modern and/or optimized versions or Quicksort as it is unclear if the claim that it becomes faster than Quicksort is correct :)

[maweki]

I think assignment into the buckets can be done in parallel.