A computer architecture in which program instructions are stored in different memory from dATa. Each type of memory is accessed via a separATe bus, allowing instructions and dATa to be fetched in parallel. Contrast: von Neumann architecture. [Why Harvard?] (2004-01-14)