null character without reason in written files

[Steve Kargl]

On Thu, Oct 31, 2013 at 09:24:30PM +0100, FX wrote:
> > I know it but that's not a RNG with good statistical properties.
> 
> The Fortran standard makes no guarantees on the statistical properties of the PRNG. Obviously, there is no single ?perfect? PRNG, so use something that checks the boxes you need checked!
> 
> But, I wanted to stress that gfortran?s implementation of RANDOM_NUMBER, as documented at http://gcc.gnu.org/onlinedocs/gfortran/RANDOM_005fNUMBER.html , relies on pretty robust and certainly well-studied PRNG: George Marsaglia's KISS. It has been so since a long time now, probably somewhere around version 4.3.
> 
> In particular, it does not show the ?obvious? telltale signs of bias.
> 

Hi FX, 

>From the description given in a different reply, it 
seems that he is seeing the distribution of floating
numbers within the interval [0,1).  I need to think
about this a bit more.

-- 
Steve