Mon Sep 27 11:00:50 CEST 2010

Values and Transitions

This places the previous post[1] in another light.

If you want to model a dynamical system (a system with internal
state), you need to have concepts of both state and transition.

For physical (Newtonian) dynamical systems this is very obvious: you
always need two state variables: position and velocity.

For discrete systems there is a clear analogy.  To be able to detect
and compute in terms of value change, one needs to have access to the
_previous_ value, so there are also two state variables: current and
previous position, or current and difference.

[1] entry://20100925-112537