Does this really make that much difference? I kind of view the hjkl vs arrow keys like I view the standard debate about motion commands vs. the mouse.
Lots of people say they are faster using motion commands, and I know I use search motions 85% of the time. But if I want to go somewhere that's visible on the screen, the mouse is almost always quicker. And I've used vi and derivatives for 20+ years now (although obviously they initially didn't have mouse support :-)).
Similar for hjkl, especially if you use single cell motion mixed with whole page motion (Fn-arrow on a macbook), I just don't see hjkl + ^D^U being quicker. It seems more like something that might be useful every once in a while, but not terribly useful 100% or even 50% of the time.
EDIT: And the book says this:
"type the number of characters or lines you want the cursor to move, followed by h, j, k, or l. This technique cannot be used with the arrow keys: left, right, up, and down."
For me the mouse is never quicker. I think that over time your brain kind of memorizes the number of lines (rows) and characters (columns) available in your screen and it certainly gets to a point where you don't really think about the science behind motions and you end up just using motions defined through past experiences. I amaze myself (and my coworkers) when I apply a seemingly random number to a motion and I end up exactly where I wanted to be. Last time a coworker asked me if I'm a math freak and I did the calculations in my head when doing a complex motion in a second, and even though I answer that I am (jokingly), I know that it's actually just a combination of muscle memory and past memories.
In any case, even though I don't agree with the school that says that to learn vi/vim you should forgo arrow keys and non vim (cmd + s to save, command + left arrow to go the left of the line, etc) key combinations, I think that a major point of inflection while learning the editor is choosing a trick/combination/command that you want to learn because it will increase your productivity and keep forcing it, even if it means not using the arrow keys or other combinations you're not used to, until it becomes second nature. And then do the next one and so on.
For me it's not about wall clock quicker, but perceived quicker. I use a couple of different keyboards throughout the day, and they all have the arrows in slightly different places (more variation than letter keys...) which makes it similar to the mouse...
Namely, the amount of concentration required to hit an arrow or get the mouse to do what I want feels slower because I have to pause my thoughts, rather than just queue up a nerve message to my fingers to "then go over there and...". Stopping the train of thought to manipulate the mouse - which requires a lot more visual feedback and focus than just knowing that the cursor will get there - makes me overall slower because I'm context switching.
It makes some difference. It is worth doing for some comfort and speed. But it isn't true that you can't get by without it. It's just much clumsier than it needs to be.
Lots of people say they are faster using motion commands, and I know I use search motions 85% of the time. But if I want to go somewhere that's visible on the screen, the mouse is almost always quicker. And I've used vi and derivatives for 20+ years now (although obviously they initially didn't have mouse support :-)).
Similar for hjkl, especially if you use single cell motion mixed with whole page motion (Fn-arrow on a macbook), I just don't see hjkl + ^D^U being quicker. It seems more like something that might be useful every once in a while, but not terribly useful 100% or even 50% of the time.
EDIT: And the book says this:
"type the number of characters or lines you want the cursor to move, followed by h, j, k, or l. This technique cannot be used with the arrow keys: left, right, up, and down."
Which is not correct.