Stored-program computer
A computer that keeps its instructions in the same memory as its data, so a new job only needs new instructions to be loaded.
A stored-program computer keeps the list of instructions it is following in the same memory as the numbers it is working on. Machines such as the wartime ENIAC were set up for each new problem by plugging cables and setting switches, which could take days. EDSAC in Cambridge, which ran its first program in May 1949, read its instructions into its mercury memory instead, so changing jobs meant loading a different program.
An everyday example is a phone: the same hardware becomes a map, a camera app or a calculator, depending on which program it loads. LEO, built by J. Lyons and Co. on EDSAC's design, switched between bakery valuations, payroll and teashop orders in this way.
Because instructions are just numbers in memory, a program can even build or change other programs. That is how compilers, operating systems and app stores work today.
A common misconception is that the first computers were all programmable in this sense. Several early electronic machines were not: they needed rewiring for every new task.
See it in action
Coming up in future stories
- The Silk Portrait Woven From 24,000 Punch Cards 1839 · Lyon, France and London