Date: 03/24/15-11:38:11 PM Z

Hi Cesar,

C0 functions have certain symmetries in their arguments. Those
are applied by PaVeReduce:

f1 = PaVe[0, 0, {0, 0, m1^2}, {0, m2^2, m2^2}] // PaVeReduce
f2 = (PaVe[0, 0, {m0^2, 0, m1^2}, {0, m2^2, m2^2}] // PaVeReduce) /.
m0 -> 0
PaVeOrder[f1 - f2]

returns 0.

Cheers,

Am 24.03.2015 um 22:46 schrieb Cesar Lattes:
> Hi,
>
> I found a very strange behaviour of the function PaVeReduce when one of the variables is much smaller than the others. It seems that I don't get the same reduce if I apply PaVeReduce before or after sending a specific variable to zero.
>
> For instance, if I try to compute:
>
> f1 = PaVe[0, 0, {0, 0, m1^2}, {0, m2^2, m2^2}] // PaVeReduce
>
> and
>
> PaVe[0, 0, {m0^2, 0, m1^2}, {0, m2^2, m2^2}] // PaVeReduce
> f2 = % /. m0 -> 0 ,
>
> then I get a result which is not zero:
>
> f1 - f2 // Simplify
>>>> 1/2 m2^2 (C0(m1^2,0,0,m2^2,0,m2^2)-C0(0,m1^2,0,m2^2,0,m2^2)) ,
>
> Am I doing something wrong? What is the good way to take this kind of limit?
>