🚀 go-pugleaf

In article <36984@sdcc12.ucsd.edu> whart@cs.ucsd.edu (Bill Hart) writes:
>I have an application for which I need to generate a random 'vector' of 
>bits of length N.  I have looked at stuff in Numerical Recipies for
>generating a random sequence of bits, but it remains unclear whether
>such a generator would be appropriate for generating a random
>vector of bits.

>I need/want to have the distribution of vectors remain 'random'.  Right
>now I'm generating a random float in [0,1> and testing to see
>if it is less than 0.5 or not.  This works, but I'm sure there is
>a better way of doing this.

>Any ideas or pointers?


I would suggest using a long-gap shift-register procedure, using
XOR or + (if full word integer addition is possible).  By long-gap,
I mean something like using the n-406 and the n-607 words to get
the n-th word.  If the appropriate length seed (here 607 words) is
something like physically generated (or even the noise from usenet
news), the individual bits should be fairly good, at least after a
few runs to hand initial bias.
-- 
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
Phone: (317)494-6054
hrubin@pop.stat.purdue.edu (Internet, bitnet)  
{purdue,pur-ee}!pop.stat!hrubin(UUCP)