You generally had no choice about it. For example, Dartmouth Time Sharing supported the Teletype ASR-33[1], which ran at 10 characters per second. So if you were printing a moderately long program you would need to wait 5 minutes or more.
Also the density of the tape wasn't very high (I'm guessing less than 10 characters per inch). Which means that 5 minutes of printing amounted to 300 inches, or 25 feet. It wasn't really a problem, because the paper tape naturally curled up.
Imagine two typewritten pages on standard A4/letter-size paper, which is ~11 inches high - that's probably the entire RAM of those systems at the time; several KB at most.
that's probably the entire RAM of those systems at the time; several KB at most
Dartmouth Time Sharing System[1] started out on relatively small computers, but by 1965 was running on a very state-of-the-art mainframe, the GE 635[2]. It had about 1 MB memory. That doesn't sound like much right now, but:
The 635 version provided interactive time-sharing
to up to nearly 300 simultaneous users in the
1970s, a very large number at the time
It's the wheel of reincarnation. What we now call "the cloud" was called "time-sharing" in the 1960s.
Also, much of the actual time-sharing control on those GE's was handled by a separate communications processor (e.g. Datanet-30) with its own memory space and registers.