There seems to be quite the influx of "the shell and terminal emulator, but better" and nobody makes it actually better. Most of the "upgrades" (since to me, t

There seems to be quite the influx of "the shell and terminal emulator, but better" and nobody makes it actually better. Most of the "upgrades" (since to me, they are not upgrades) revolve around making it pretty or constantly showing irrelevant data or "helpers" on screen. The downside and upside of a shell is that you need to have a mental map of what you want to do before you can do it. That is not how the current generation seems to approach technology at the foundational level anymore and thus the 'need' for making it pretty and shove potential 'answers' to what you wanted to do in your face is what we get from the 'new' incarnations. I have no idea if it would actually be better or if it is a discoverability or usability issue, but after learning how to use a shell I personally never found the need for a shell or terminal emulator that does more than accept my commands and show the results. To me, it's almost like it needs to be as close to a system monitor as possible while still allowing high-level interaction with the operating system. Not an 'environment' separate from the system. It's the same with powershell. I get that it is more of a mini-os inside the os with its own things and data structures passing between commands (which sometimes are actual commands, sometimes are not... and I'm not talking about shell bultins), but if I wanted a REPL I would choose to use a REPL. (but one could argue that a POSIX shell is just as much a REPL shell) Creating a terminal emulator that 'connects to SSH' is of a similar issue to me. The SSH-client connects to SSH, not the terminal emulator. And the SSH-client runs as a program in the shell in the terminal emulator. You could have some sort of weird GUI-shortcut to open a terminal emulator window or tab, which starts a shell and an SSH-client inside that shell, but what is the point of that if you are going to open a shell anyway? You're essentially adding an extra step (the mouse) where there was none. And now, instead of having the model in your mind where you think "I am going to use an ssh client to connect to an ssh server on a system somewhere" you have the "I click some buttons, fill in a form, and now I am 'on' the other system somehow". The first part is portable as you can swap out the program with another program and your model works, where the second one isn't and you essentially end up remembering waste for every single command that is GUI-fied and you have to think about them in pictures and locations instead of commands and goals. And now this rant is over before I make it worse.