[Help-gsl] turn MC integrator into sampler?

[Shao-Feng Ge]

Hi,

Is there any possibility of turning the GSL monte carlo integrator, such 
as Vegas and Miser, into a random sampler?

What I am naively thinking is, during monte carlo integration random 
points are generated and the function values are calculated at these 
points. The result of the integration is a weighted sum of these function 
values. By binning these function values into a histogram, the integration 
code can serve as integrator and sampler simultaneously. This can be a 
very useful tool.

The only thing that I'm not sure is which parameter in the gsl code carry 
the info of each sampled point's weight and how to deliver it to the 
user-defined function, such that each time the user-defined function is 
called, we can register the point and the weight associated with it.

See if your guys have any good idea.

Shao-Feng