Changes between Initial Version and Version 1 of Ticket #17123, comment 50


Ignore:
Timestamp:
07/29/17 19:18:59 (4 years ago)
Author:
darij
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #17123, comment 50

    initial v1  
    1515You should be comparing with `binomial(n-1, n-k)`. These are really partitions of n-k you are summing over (the first part is just for convenience, as you clamp it to k).
    1616
    17 Sorry, but it makes no fucking sense to me to diverge from the (n choose k) = n(n-1)...(n-k+1)/k! standard for negative n as long as k >= 0. The binomial coefficients (n choose k) for a fixed k are a polynomial in n when n is nonnegative; why on earth would you want to break that?
     17Sorry, but it makes no sense on earth, in heaven and in other places to diverge from the (n choose k) = n(n-1)...(n-k+1)/k! standard for negative n as long as k >= 0. The binomial coefficients (n choose k) for a fixed k are a polynomial in n when n is nonnegative; why on earth would you want to break that?