I know that GUI does not cover most of functionalities, for good reasons - being specialized to task (like files app), it provides more fine-grained experience.
Yet, I find that there are common commands which is terminal-only, or not faithfully implemented. for instance,
- Commands like
apt update/apt upgrademight be needed, as GUI may not allow enough interactions with it. - I heard some immutable distros require running commands for rollbacks.
These could cause some annoyance for those who want to avoid terminal unless necessary (including me). Hence, I bet there are terminal emulators which restricts what commands you could run, and above all, present them as buttons. This will make you recall the commonly used commands, and run them accordingly. Is there projects similar to what I describe? Thanks!
Coming back to this with thoughts. What you’re describing sounds a lot like a menu tree.
“Press 1 to do this, 2 to do that, 3 to go to submenu A, 4 for B,” etc.
1“You have pressed 1. Do you want to turn on option ABC? [Y / n]”
Y“Do you want option QWERTY47? [Y/n]”
N“Are you sure you want to run
notthebees --abc --no-qwerty47? [Y/n]”N“Aborted.”
It sounds like a standards problem waiting to happen because no two menus will be alike, but hey, things like this can and do exist, and setting one up isn’t that hard, only time consuming.
This was my thought exactly. And I also had the same assessment that having various arguments be context aware would be challenging, since some have sub-arguments of their own (with further sub-arguments of their own, etc.) but can sometimes be strung together all on the same line. How do you determine if someone wants an ascendant argument or a descendant argument when you’re three layers deep into the tree?
You would have to make opinionated decisions, which was the whole reason to avoid scripts in the first place. Seems like it would be better to just make executable scripts (which is what Fedora Atomics basically do with the
justcommand) or gamify learning how to work in the terminal.Something like this can kind of be achieved programmatically by unraveling bash completion arguments and loosely parsing terminal help strings.
They aren’t all formatted uniformly though, so you’ll need to come up with a filtering mechanism to prevent returning garbage. You’ll also always be a little out of date…