Since the number of bounces could be infinite I wonder what kind of termination mechanisms is used here. It seems like both probabilistic approach and fixed depth would work in this context.

I understand how we randomly sample the direction and distance of bounces. But what is the point of connecting each shading point to the light?

I think the idea is that we are trying to represent how light bounces around through participating media such as fog. So we consider how light hits those points and how they bounce light to the next in order to model how different parts of it are lit

Is the distance we use between points based on the mean free path of photons in that substance? Is there a difference in the mean free path for scattering vs absorption? http://glossary.ametsoc.org/wiki/Mean_free_path

I'm kinda confused about how this ray is being traced through the medium. From out perspective, are we tracing to an intersection at the medium, then doing the random walk with intermittent points connecting to a light source inside the medium? Where is the light source established, since I thought the phase function isn't constant and the scattering inside won't be uniform.

Im curious how much of an effect participating media has on increasing the computational complexity of rendering. Intuitively, it feels like taking into account scattering from water or clouds requires a lot of effort.