The summer of 1978, I spend some time to convert a large Fortran program in the IBM dialect to Fortran on a Cyber mainframe. The program consisted of about 1500 punch cards. At first I would load the whole deck every time. After some time, I discovered it was possible to store the program on disk and edit them by-line using a program called Update. This still requires typing punch cards. Everytime, I checked the cards many times to make sure, I did not make any mistakes. And then it was waiting before the monitors showing he input, the execution, and the output queue, If it was out of the output queue, you still had to wait before the output was dropped in one of the labled boxes, which could take another ten minutes. In those times memory usages was billed in the Kbytes per second. I did it for nothing. Just the fun to work on a real mainframe was enough. Afterwards, I was rewarded with the book `Finite Mathematics' by Seymour Lipschutz.
The person giving me the assignment also wrote programs in some kind of simulation language where the lines could be in any order. Sometimes he would shuffle the cards while standing in line for the cards to be read, just to make fun of the other waiting.