In 1963, my second programming job out of school, I worked for CDC in Palo Alto. My group was developing a COBOL compiler for the 3600, but we did not have a 3600. MSU did have one and was not using it from 8pm to 5am. So CDC shipped the group to East Lansing where we worked the overnight shift. At the time about half of us were women, all single. (That has changed; there seem to be fewer women now, and many are married, or were.)
Of course the compiler was coded onto punch cards, which occupied several boxes. Those cards were saved onto tapes. We made corrections with punched correction cards keyed to numbered lines in the matrix printed listings of our code cards. The cards were still punched by a staff of keypunchers. From time to time, my corrections would blow up because part of the deck had been put in upside down.
At Purdue in 1968 my husband did his thesis programming on their brand new 6600. We would run his programs during the "off hours."
The school had a massive card reader for student's programs. I warned my freshmen students they would have time to leave the building before their decks got processed.
Unnumbered FORTRAN, anyone?