caps complex tests



On 11/14/2010 4:59 PM, Richard Hennessy wrote:
> Speaking of interpretation. radcan(sqrt(y^2-2*y+1)) is using 
> sqrt(x^2)=x so you get y-1.  Well, why not 1-y instead as your example 
> below points out?
>
> radcan(sqrt(y^2-2*y+1));
> -> y-1
>
There is one answer which corresponds to PRI (positive real 
interpretation), which is y-1.
That is what radcan is supposed to return.

It does NOT return 1-y   for sqrt((1-y)^2).

because that is not the PRI.     y-1 is.

RJF