Conservation of information: applications in functional, reversible, and quantum computing

Abstract

Microscopic physics is entirely reversible, yet almost all computation is done using irreversible processes. Adding two numbers together, for example, destroys information, unless one of the numbers is retained. Over the past few years, some advantages of taking reversible computation seriously have emerged -- first in the context of saving energy, through the development of dissipationless logic, and then in the promise of quantum computation, which is necessarily also reversible and dissipationless.The necessary conservation of information in these systems raises the question of how we can best capture this strong constraint from an abstract point of view. Already ideas such as linear logic can be applied, but the full implementation of conservative logic ideas will require language and architectural innovation and a rethink of our computational models.