MUDs and the ability to reload new server without dropping connections (a “hotboot”)…
https://boston.conman.org/2025/02/11.2
…reminded me of an old RSX-11M application from an aeon ago, running part of a state government.
The app was too big for the PDP-11 (a common problem, as 32KW wasn't all that much even then), so the app removed the operating system from the computer, and ran, well, standalone.
Yes, you could do that back then.
That all worked swimmingly until somebody pressed ^C control-C on their terminal session, and the terminal driver then trapped into, well, nothingness.
Since the app code was too big for the PDP-11 it was running on, that smaller PDP-11 was shut down, the bigger local PDP-11 was switched over and re-booted and the app code loaded, and the bigger PDP-11 was then powered down.
The operator then pulled the core memory out of the bigger PDP-11, walked it over to the smaller PDP-11 and plugged it into the backplane, toggled in the address of the app's main loop onto the front panel switches, and toggled “go”, and off the smaller PDP-11 went running its too-big app.
Yeah, you could do a simple form of app checkpoint-restart with core memory, given core was persistent. And yeah, a bigger PDP-11 would have been helpful.
(Yes, SC was a creative administrator of those PDP-11 boxes. Swapping core never would have occurred to me.)