ABSTRACT
Abstract This personal essay recounts a misadventure in attempting to find a quantum algorithm with a “general purpose” flavor. Exact arithmetic with large integers can be conducted modulo a list of relatively prime numbers, the answer being reconstructed in conventional terms only at the last moment. Implementing this process on a molecular level would lead not to a quantum computation, but to a massively parallel classical computation. It has been observed that something similar is going on, in practice though not in principle, in the nuclear magnetic resonance implementations of the Grover search algorithm. Certain other NMR computations, which are frankly of an “ensemble” nature, are even closer to my proposal, though more practical. Comparing and contrasting these situations may cast light on the current semantic disputes over what constitutes a “quantum” computation.
