New computer architectures and their relationship to physics or why computer science is no good

• DOI: 10.1007/bf01857728
• Publication Date: 2005-06-30T18:14:11Z
• Authors: W. Daniel Hillis
• Keywords: 0103 physical sciences; 01 natural sciences

In the past, computation scientists have found it convenient and productive to adopt a model of the computational universe that was very different from our models of the physical universe. This is chaning. As we build bigger computers out of smaller components, our models of computation are forced to change. There is reason to hope that our new models, for specific systems, will be similar to the models of physics. The paper is divided into three sections. The first argues that computer science is missing many of the things that make the laws of physics so powerful—locality, symmetry, invariance of scale. This is why physics is so nice and computer science isn't. The second section gives an example of a new-wave computing machine, and shows some physicslike laws that apply to its computations. The final section gives some reasons for expecting this convergence of physical and computational law.