> vi is everywhere.

Although, surprisingly, most command-line interfaces and command-driven apps (like redis client and psql) use emacs key bindings for command history.

Most command-line interfaces use the readline library. To switch them to vi bindings put the following in ~/.inputrc:

  set editing-mode vi

This isn't a deliberate choice though on their part, is it? Isn't it just a consequence of everyone depending on readline and readline using emacs-style key bindings?

Some Emacs-style bindings, like ^W (erase word) and ^U (erase line), are baked into the UNIX tty discipline -- basic implementations of these bindings exist within the kernel.

Every machine I’ve ever ssh’d into has nano. Maybe that’s a new thing (most of my DevOps work has been in the past 4-5 years) but nano feels great for the kind of work I do when I ssh into something. If I need to change a config file here or a script there, nano can do it. If I need to write a whole new complicated script, I’m gonna do it on my local computer (in WSL or my Linux partition) using VS Code and deploy it to the server so it doesn’t get wiped by accident when I terminate a VM.

If you work with embedded linux, nano is almost never there.