As far as we know, there's no way to exactly simulate physics on a digital computer. So any simulation would be a numerical approximation. If we're in such a un

As far as we know, there's no way to exactly simulate physics on a digital computer. So any simulation would be a numerical approximation. If we're in such a universe, it would have consequences for understanding physics. For instance, we might be able to detect numerical round-off error with some sufficiently precise experiment. Back to your analogy, if you divide 42/9, a calculator will say something like 4.66666667, which is only an approximation to the true answer. So it makes a difference what substrate the number exists in. If we are in a simulation and could figure out exactly how the approximation is done, we might find some useful tricks to do with it. Most practical physics simulators we have today (say for simulating robots) can be triggered into producing wildly wrong results, where energy isn't conserved. That would be cool.