[NatCh]

Originally Posted by HarryT

Sorry, I don't understand. Take a typical application, such as a web browser, a word processor, or a spreadsheet. Explain to me how it would be "more powerful" to have to type a command to follow a hyperlink on a web page, rather than simply clicking on it? Or how is it "more powerful" to use a CLI to edit a paragraph in a word processor, rather than just visually scrolling to it, and typing it?

Okay, I think I can help with the concept here. You're actually looking right at it, but you're not 'looking right.'

So, to take the concept completely away from computers where it might be seen in sharper relief, I ask you to consider Public Address equipment, loudspeakers, I mean.

I run sound-boards at church, and have for a couple of decades now. It's not the complex business that you'd get into at a recording studio, or radio station, but we have a sound-board with lots of knobs and sliders, so it's somewhat complex.

Originally, all we had were a few microphones, some speakers in the wall and a single box amplifier with about 2 knobs on it: power and volume.

These days we have a board with about 12 channels each of which has a slider and a dozen knobs for various functions, plus an separate amplifier, that more or less has the same two controls, power and volume, plus the speakers and microphones ( and tape decks, and recording equipment, and CD, and DVD, etc., etc.), of course.

Both set-ups do the same thing, but the second, more complex one, allows greater control of the operation. If person A's mike sounds a bit tinny, I can tweak the equalizer settings for that channel and fix it. If person B sounds too soft, I can turn just his mike a bit louder, and fix it. I couldn't do either of those things with the old "One Box" system.

But on the other hand, the One Box's learning curve was a short, gentle slope, and the soundboard's is a towering cliff, if you're coming at it cold. But, that very complexity is what gives us greater control over the end result. It's a more powerful interface specifically because it's more complex. That's what makes the greater complexity worth the bother.

Now, to put it back into the original context, the GUI makes it much easier to do the things that it covers, the learning curve on them is quite gentle, but the GUI does only the things it's programmed to do in the way it's programmed to do them -- if you don't like the results, you're still stuck with them. The very fact that the CLI tries to do less for you means that it can offer more flexibility in how whatever you're doing gets done. Yes it takes a lot more time and effort to get to where you can do anything of any significance, but once you do, you have a lot more control over the final result. Sometimes that's a bad thing, of course.

That's all generally speaking, of course, I do realize that there are examples of CLI's that offer no flexibility, and GUI's that offer tuns of it, but I'm just trying to highlight how greater complexity can offer a more powerful interface as a trade-off for that complexity.

I said that you were looking at the solution, but not looking at it right: what I meant was the keyboard mouse thing: the mouse is a lot easier to use for navigation, and in many cases control, but if you had only a nice, elegant, simple mouse, data entry would be a complete nightmare. The keyboard's greater complexity allows you a more powerful data entry mechanism. Conversely, the keyboard's very complexity prompted the invention of the mouse because navigation is kind of a pain on a keyboard, it has a steep learning curve for that sort of thing.

Originally Posted by HarryT

Compare the task of writing a letter in "Word" compared with doing the same thing in "vi". Which is the better tool for the job?

This comment is sooo close to the point I'm making, but just a bit off. The question of GUI vs. CLI isn't so much whether you'd prefer to write a letter in Word or vi, but rather would you prefer to write a letter with only your mouse or only your keyboard?