simpler version of MWC256?

I found a different (simpler) version of MWC256 that George Marsaglia himself posted on the internet. It has the same 2^8222 period and Marsaglia vetted the quality of the PRNG as well. The version that we currently use comes from Feb 25, 2003 https://groups.google.com/d/msg/sci.math/k3kVM8KwR-s/jxPdZl8XWZkJ:

```
static unsigned long Q[256],c=362436;
unsigned long MWC256(void){
unsigned long long t,a=1540315826LL;
unsigned long x;
static unsigned char i=255;
t=a*Q[++i]+c; c=(t>>32);
x=t+c; if(x>32);
return(Q[i]=t); }
```

Why was the first version chosen for mwc-random?

Read the two George Marsaglia references included in the issue and compare them with the current MWC256 implementation in mwc-random. Establish why the existing version was selected and whether the simpler variant is suitable; done requires a documented decision, not merely a code change.

Written by the indexing model from the issue text.