The Apollo Guidance Computer used core rope memory; the software was literally woven by hand into it.
One of the most common anecdotes to illustrate the growth of computing power over the 20th century is to compare the computers on board the Apollo moon missions with, say, a modern calculator. Certainly the cellphone in your pocket (assuming you own a cellphone, or a pocket) has more memory than all the moon missions put together. But was your cellphone handmade?
Core rope memory was used for the Apollo Guidance Computer, the computer that ran all the navigation and controlled the spacecraft itself. For such a crucial function you needed as much reliability as possible. Because space (inside) was extremely limited, it needed to be extremely compact. And you really really did not want to accidentally wipe or overwrite it.
To meet all these criteria, the software for the guidance computer was literally hand-woven into it. The memory looked rope-like, hence the name, but really it consisted of clusters of wires (“ropes”) wrapped around magnetic cores. In contrast to normal computer memory, it wasn’t the cores but the configuration of the wires that carried all the information. You couldn’t change anything stored in core rope memory without moving the wires around; in other words, you would have to physically rebuild the memory to alter the software. This memory, in addition to being very reliable, was also a very efficient use of space: 2.5MB per cubic metre, unbelievably dense by the standards of the time.
So who had to do the painstaking memory weaving? It should not surprise you to know that it was women doing the hard work, hence the alternative name “Little Old Lady memory.” And the software held in the core rope memory was programmed by a woman too: Margaret Hamilton was the lead Apollo flight software designer.