CNDTIC TABD- ~ U.oa-mouliced Justifica,i....... By

attempted to survey the current status of the reversible logic approach to digital computing. Our aim was to form an opinion on which future developments of this subject might be relevant to DARPA's computer-related interests and possibly worthy of its support. We consulted one of the pioneers in this subject, Dr. Edward Fredkin of MIT, at some length and brought ourselves up to date on the (not very extensive) literature. Our basic conclusion is that the importance of reversible logic depends crucially on the physical architecture of the computer: It is irrelevant to the current scheme in which packets of charge are stored on, and moved between, structures of order one light wavelength in size, but might be relevant and even essential if the basic information-handling units were of molecular or atomic size (a distant but not necessarily-unattainable goal). The question of physical realization of reversible logic elements has been almost completely neglected * in favor of the abstract questions of how, given the existence of reversible logic elements, one could wire them up to make a useful computer and how one would program it. We think that *Apart from some interesting "existence proof " work ot Fredkin et al.