3. The computer science approach

Do not become attached to the things you like, do not maintain aversion to the things you dislike. Sorrow, fear and bondage come from one's likes and dislikes.

— Buddha

The computer science approach to modelling physics with CA is qualitatively different from ei­ther theoretical or experimental physics, or from the kinds of abstract mathematical work that so often leads to progress in physics. The problem is that the study of cellular automata is both a theo­retical and an experimental science. However, the experiments, which often produce results we did not anticipate, are not like physics experiments. They are the kind of experiments that never ex­isted before the age of the computer. While there are not yet any formal methods for creating CA models with properties one is looking for, practi­tioners of this arcane field have developed skills over the years that make searches more efficient. The primary skill is the ability to understand the kinds of behaviors that one can expect to find as a consequence of certain kinds of CA rules, structures and initial conditions. In suggesting CA models for DM we are guided by both basic prin­ciples and aesthetic considerations. It is only aes­thetics (including economy and simplicity) and initial conditions that separates one UCA model from another.