I'm not sure I follow why the density function can't be constant over the interval any more than it can be the usual slanted bell shape it is? I'm more than fine that the actual light is sampled out of this in quantized bins and amounts, that shouldn't be any different than any other light.