3 ms·
I'd say that "on average" results may be very misleading, in formal statistics and in results like this. Variance and median are very important: see comments ab
by nine_k 7y ago
I'd say that "on average" results may be very misleading, in formal statistics and in results like this. Variance and median are very important: see comments about Haskell downthread. (Ideally you have a histogram.)
In particular, some languages with a good average may contain pitfalls that could lead to abysmal outlier results (e.g. C++ or Scala, in entirely different ways), some would keep you from doing really dangerous things but also would not let you achieve spectacular things (e.g. Go).