Re: [Bug-gsl] inaccurate tabulated values in gsl-1.11/randist/gausszig.c

[Joseph Pang]

Hi Brian,

After reading the code more carefully, I have to retract my earlier comments.  
The author is using an upper exponential bound for the tail of the Gaussian.  I 
was mistaken in thinking he is using an exponential approximation for the 
Gaussian tail.  In that case, it is the right thing to use the exponential tail 
in the Ziggurat pre-calculations.

I apologize for the false alarm.

Regards.

Joseph Pang