Although not a PDP-8 (introduced 1965), I did complete my university study of operating systems by working on a PDP-11 (introduced 1970). Because it was an operating system course, we couldn't just turn the thing on and go. We had to bootstrap the hardware by hand, entering a few instructions by toggle switches on the front of the machine. That allowed us to download our hand-built assembly language operating system and try to get it to output messages to an attached dot matrix printer terminal. Real primitive stuff.
Note that for the mainstream programming stuff (e.g. FORTRAN, C, PASCAL, etc), we started out writing our code on sheets of paper. Then we'd go to the basement of the computing center and type them on punched card machines to produce a deck of cards with our program, ready to go. Then we'd turn that deck of cards over to an operator who would run it through a card reader about the size of a refrigerator on its side. Then we'd get the printouts of our results a couple hours later after it eventually got run on the university's timesharing UNIVAC 1108 mainframe.
Here's a guy explaining the operation of the very model of punched card machine that I used. I find it oddly depressing that I worked on what is now a museum piece.
Imagine a room with 20 of those things clattering away.
Things got more advanced when we were allowed to use the dot matrix printer terminals directly. I'd include a video, but they're invariably blurry and awful. Just imagine something printing at 300 baud.
Eventually, the university received new minicomputers and "glass teletypes" running at 9600bps. At 300bps, a line would take about 2 seconds to print. At 9600bps, lines would just scroll right by on the screen. It was amazing to us.
And, of course, all the keyboards were mechanical, with deep throws and a loud clack.
Compared to that technology, everything we see now is on steroids. Except perhaps the keyboards. I feel rather honored to have been around to experience those times.