For the new "Uniform" class in Issue #203, create optimized "simulate" method ...

[llorracc]

... that just reshuffles the order of the points in a uniform distribution (which can then be mapped into the corresponding points in an equiprobable discretization of any distribution.

Related to #203

[albop]

There could be two ways to do it:

• write a .simulate(... method='stochastic') that would use its own implementation only if method=='snowglobe' and would defer to the ancestor method return super().simulate(pass all arguments)
• write a new .snowglobe(...) method
  Here, snowglobe is just a placeholder a the method name until we find the right one.