Question: Strange interaction between add and random numbers... a curiosity?

 

As I was comparing visually the first terms of a priori identical sums produced by add , I was surprised to find them different.
So I suspected some error in what I have done, until I realized that add randomly permuted the terms.
Each term is of the form (R + P)2 where R is a random number and P a polynomial.

This behaviour is illustrated in worksheet add_changes_ranks.mw and appears only when random numbers are used (provided the seed is not forced to some constant value)

Does someone ever onserved that or have any idea of what happens here (maybe this behaviour no longer happens in recent versions?) ?

Thanks in advance

Please Wait...