What is the use case of RandomCover?

[Ivan Kazmenko]

monarch_dodra wrote:
> In theory, there should be no reproducibility breakage, as the 
> generators themselves will not be changed (they are _strictly_ 
> modeled after boosts' (the ones we have anyways)).

Indeed, I just looked at Boost and C++11 implementations, and 
they currently have these divisions too.  Having this kind of 
reproducibility across languages is a benefit in itself.

bearophile wrote:
> Is this to report in Bugzilla?

Sorry, I fear it's a false alarm, and the answer is no.  In any 
case, before advocating such a change further, I would like to 
know why Boost and C++11 people didn't do that in their realms.  
Maybe the division is actually more portable.  Maybe extracting 
bits is indeed faster with MT19937, but it can produce poorer 
quality random values with RNGs which are not so well-distributed 
bit-by-bit.

-----
Ivan Kazmenko.