Just a small comment: We do not have random functions in our computers. We have psuedorandom numbers in our computers.
Also, I think a lot of what you argue can be resolved by a careful examination of measurement theory and real analysis (see for example, lebsegue theory).