On 7/26/09, Barton Willis <willisb at unk.edu> wrote:
> With CVS Maxima, the results are different:
>
> (%i18) sum(-k, k, -a, -1), simpsum, factor;
> (%o18) a*(a+1)/2
OK, that's good news, glad to see the bug was fixed somehow.
> Manifestly empty sums simplify to zero:
>
> (%i19) sum(-k,k,5,-1);
> (%o19) 0
>
> I suggest that we consider restoring the sumhack option.
OK by me to restore sumhack. (I'm the one who cut it out in a
larger revision of the summation code in 2005. I think I was
acting on a suggestion from someone else, although I can't
find a record of it at the moment.)
If we want to be very careful, the symbolic result should respect
sumhack as well. I don't know if it's worth the trouble.
best
Robert Dodier