...To conclude, let me mention briefly one nonelectronic analogue computer from the early 1950s. The Phillips machine was a hydraulic simulator of a national economy. It wasn’t particularly accurate by modern standards (about 5%) – but, as mentioned above in the context of electronic analogue computers, accuracy was not necessarily the prime concern. In the Phillips machine monetary flows (income, tax, even foreign exchanges) were represented by flows of coloured water. Up to seventh-order systems could be simulated.
Seeing the machine working is different from pictures of it, as those who have seen the machine working readily attest. Spectators could not only see the red water streaming through the pipes, but also hear the bubbling and splashing as it ran through the machine. They were able to see not a 2-D picture or system of equations, or even a static 3-D representation, but the kind of interrelated and dynamic causeeffect changes over time that economists suppose to happen in the circular flow of the aggregate economy. The working machine was a 4-D representation.