It's interesting how linux kernel dumps work.
You actually have a linux in your linux, and when linux fails, linux takes over and can dump the linux address space so when linux gives up and boots linux, you can use linux to do an autopsy on itself. :)
Or, in other words, you allocate a reserved address space and load a linux#2 kernel in it. When you encounter a panic condition, the system can jump to linux#2 in the reserved address space. It will be non-corrupted and can dump the address space of linux#1 to disk (or a remote system) and reboot.
kexec is the basis for all this, and it can do a lot of other interesting stuff like upgrade in place.
Yo dawg, I heard you like Linux...
Side note, running with your root filesystem on a ZIP disk is not performant.
It's great to see Linux installs going this direction. The ease of install definitely made switching to BeOS simple back in the day.
Plus BeOS 5 had both Hotline (https://en.wikipedia.org/wiki/Hotline_Communications) and IRC clients.
In the same review, Scot went on to say: "I tested the Windows installer on several machines, and was amazed. Not only did everything work as advertised -- it worked better than I dreamed possible. BeOS has always been the easiest OS on the planet to install, bar none, but this is ridiculous. BeOS is up and running in five minutes or less, and everything is totally painless. Linux distributors should be taking notes." Sounds like someone did! ;-)
[1] https://birdhouse.org/beos/byte/13-beos_5/