An example: the atomic field laboratory
the simulation of physical systems
When you build a model, or a computer simulation, of something you consider to be 'physical', you have in fact just taken a kind of set of notes for yourself. They are for you to read, and there is nothing of what's on the page that has any necessary relationship to the world outside of your mind. You have created an idea, and an external artifact to remind you of the idea.
But this isn't good enough, of course. You need to make the model more comprehensible to other people. You need to do the extraordinary work of trying to make your model express enlightening principles, uncovered by examining, with great effort, some subset of the world outside of your mind.
None of this is automatic. Science is anti-intuitive, by nature, and yet it needs to, in some way, satisfy out intuition, even if that satisfaction is known to be tentative, or if our intuition is known to be incomplete.
So I want to introduce a new approach to computer simulation on these pages. If you are modeling the external world, model your own perceptual and conceptual mechanism too. Because just as much mystery lies in investigating that. These are your hidden assumptions about the outside world. They are inside you. You need to uncover them. It's part of finding out what is actually going on in the outside world.
One of the positive things about simulation, is that some of the vague assumptions that could become more explicit, do become so. This allows them, at the very least, to become something like assertions about the real world. Universally these assumptions are incorrect, but some are better than others, and some are more intelligible and enlightening than others. So they become open to criticism, which is positive.
But there are several negative things about this:
1. You may be programming surface behavior that pleases your, and other humans, non-explicit notions. Often we are either circumventing these notions or cutting them down to some simple technical term that does not capture what little we kinda know about the physical system. Worse, we may use these vague notions to bridge the gap between the results of the simulation and the effects observed in reality.
2. A corollary: we become confused about the distance between our models and reality. If some model is enlightening in one way, it might be obscuring in another.
I'm going to investigate this issue by revisiting a complex simulation of power grids, and another about cities. I'm going to try to make everything I know explicit, and in some way incorporate this, and possible interventions, visible, and parameterizable. (My 2011, cognition-free version of the power grid simulation,
RTDA, was advertised back then as a
smart grid product). As I produce them, I will link to them here.
Greg Bryant