I've actually used one of those IBM card key punches back in college. We would have to write Fortran code and type out those cards, then run the cards through a compiler, get a print out and double check our code. If everything looked OK, then we could log into a sub-frame computer (tiny - about 10X10 monotone screen) and submit our code to the main computer. Then we would hike all the way across campus to the computer science building and get the results. If the results were wrong you got to do it all over again. The next year we got some Macs, but only about 10 Macs for the whole engineering department. The irony is that even then you could buy a calculator like a Texas Instrument or a Sharp that could almost as much as the mainframe - though they had some neat Evans & Sutherland graphics stuff up in the computer science department. The whole card punch/Fortran thing was one step ahead of a slide rule, but not by much.
