Unfortunately, for the scenario I was replying to, a lot of the times when you’re doing support, you can’t see the user’s screen and are limited to verbal communication, so verifying what they typed or the output can lead to just as many problems. Any support scenario where you’re talking the other person through a series of tasks will be very dependent on how familiar each person is with the task you’re supporting. And no one Rs TFM these days, if you even get one.
Just try to verbally describe the difference between a bracket, brace, or parenthesis, or forward slash vs backslash. I’m sure it will be fine. But absolutely, a text-based interface is easier to describe in text.
Oh, I have. Now imagine giving those people a command line.
As someone who read at least 2/3 of the DOS 6 manual when it came out, and have used a variety of Linux flavors as well, a command prompt is the least helpful interface devised. What do you type there? How do you let the computer know when you’re done typing? If the answers seem obvious to you, think about why, and what on the screen would point you that way if you hadn’t had training. People are very visual, in general, and a simple interface such as a mouse that directs focus and has a minimal amount of interaction options is far easier to get started with, especially if the GUI has culturally intuitive icons (save needs updating).
I don’t think the power of the command line, or text interfaces in general, can be overstated, but even the most helpful text interfaces, such as those found in some IDEs, require prior knowledge to be useful. This isn’t going to work for the majority of people.
Melted permafrost doesn’t turn into valuable farmland. It will also cause higher CO2 emissions.