Distributed computing

Many of the basic issues about the progress of time in a universe consisting of many separate elements have analogs in the progress of computations that are distributed across many separate computing elements. In practice, such computations are most often done by requiring explicit synchronization of all elements at appropriate points, and implementing this using a mechanism that is outside of the computation. But more theoretical investigations of formal concurrent systems, temporal logics, dataflow systems, Petri nets and so on have led to ideas about distributed computing that are somewhat closer to the ones I discuss here for the universe. And, as it happens, in the mid-1980s I tried hard, though at the time without much success, to use updating rules for networks as the basis for a new kind of programming language intended for massively parallel computers.