In describing the behaviour of my example computer, the details of the hardware are irrelevant. It's only necessary to say how it pushes the numbers around. The same behaviour could have been implemented on one of many types of chip, or mechanically, or by fluidics, or (as I did) by a person simulating it. This gives rise to the idea of hardware/software independence, and to the idea that one can explain the operation of computational systems in information-processing terms, without reference to the hardware.