I “rm -rf /“ ed my Linux box in 1996 and realized what I had done about 30 seconds in. It was through like a-d in /bin. ‘Recovery’ wasn’t a strong concept in L

I “rm -rf /“ ed my Linux box in 1996 and realized what I had done about 30 seconds in. It was through like a-d in /bin. ‘Recovery’ wasn’t a strong concept in Linux os installers then, and I wasn’t sure my home directory would survive a reinstall, (not on a separate partition) and I didn’t think I’d be able to reboot successfully in any event. My brother ran the same distribution but at a school across the country; I was able to recover by carefully pulling what I needed down with ftp to his dorm computers IP. I’m not sure scp even existed then on Linux; I guess maybe we could have used netcat in a pinch. Well, this was already a pinch. In a real pickle. This was one of the first moments where high bandwidth connections to endpoints really saved me / impacted me, and I haven’t forgotten. For many years after university this kind of direct high bandwidth connection got much harder to achieve, first because we were back in a low bandwidth residential world, then because ipv4 was mostly denied to consumers, then because of nat. Today this would again be achievable, but with vastly more complexity. For a home computer rescue you’d want Tailscale in both sides. And it’s extremely unlikely you’d be using the same distribution as your sib, much less have the same libc linking. And god help you if you had to restore your systemd directory by hand.