So how could one go about creating a model of our planet's climate, or (even more daunting) the universe itself? What exactly is a model anyway?
We can start with an apparently trivial observation. The world we inhabit is complicated. Not only does it contain trillions and trillions of things, but these things also interact with each other in myriad relationships.
Why would a perfectly accurate digital model of the world not be possible?
First, because our world is analogue and the digital realm is discrete. But also because the world is simply too varied and complex. We would need a computer as large as the world itself.
So, every model has to be a simplification – computer scientists often call it an abstraction. The model builder must include only the features that have a direct bearing on the system being modelled – ignoring everything else. However, once the relevant elements have been identified, then our trusty tools of sampling and quantisation can be brought into play. To illustrate, let's look at our two examples again.