Randomness and reseeding?

[Linas Vepstas]

On Tue, Dec 12, 2006 at 10:58:45AM -0800, Susan Margulies wrote:
> then seeding it with a read from /dev/random/ using some code gotten 

Doing so will make your computations unreproducible. If you are
doing scientific research, this is bad: you might see something strange
happen on one run, and you will never be able to reproduce it again.

The only time you want to use true randomness is when generating secret
keys for encryption. 

> However, I began to notice that as I did more and more trials (like a 
> million), it seemed as if I was loosing the randomess. 

How did you notice this?  "millions" is a small number, most prng's (I
don't know gmp's) have been tested to death for such small values.

--linas