Some of us wrote our programs out carefully on paper, debugged them as well as we could by stepping through them mentally (and making register annotations, etc.

Some of us wrote our programs out carefully on paper, debugged them as well as we could by stepping through them mentally (and making register annotations, etc., on paper when required), then transcribed them to electrographic "punch" cards for the school's batch run on the local university's mini/mainframe once a month. Having a computer makes programming more convenient and less theoretical (bugs happen occasionally despite correct code), but it's not essential to learning. It's a lot less likely to happen in that way in the current environment (with ubiquitous PCs), but if we ever get to the point where PCs (or their equivalent) aren't everywhere anymore, the interested geeks will find a way -- just like we did in the old days.