The other day, I had lunch with my friend Ori Barbut and it turns out, we both seem to have a passion for a fairly mouse-free computer experience. However, the center pieces of our interaction with a computer are quite different; Ori uses emacs and I use vim. Ori is certainly interested in vim and I have quite extensively used emacs in the past, so much of the conversation circled around the why-and-why-not of the editors we are using.
I couldn't imagine writing any text on a computer without vim. Yet, the conversation with Ori made me think. Ori's environment can effortlessly do all kinds of things that sound really interesting.
It took me a couple of days to be able to express what exactly it is that makes vim stand out for me: Motion (and Text objects). When learning vim, most people focus on using the hjkl keys to move the cursor around and they try to avoid the arrow keys. You often hear people praise the fact that by using hjkl navigation, you never have to take your hands of the "main" keyboard. Although that's true, it isn't the main part for me. Vim offers a great number of ways to navigate and refer to content in semantic chunks rather than line by line or character by character. The keywords here are motion (moving around) and text objects (referring to entities). I believe that these two are what makes vim feel like an extension of my body rather than a tool that I choose to use or not.
When using vim (at a non-trivial level), you typically construct commands that consist of three components
[count]<action><motion or text object>
For example, you might type 4gk to (g)o to the location (4) lines up (k). (In fact, in this case, you could omit the "g".)
Similarly, you could type 4dk to (d)elete from the current position to 4 lines up.
So the count specifies how often you want to do something (if omitted we do it once), the action specifies what you want to do and the motion or text object specifies the target of your action — on what you want to apply your action.
When we think about how we want to edit a file, we typically think in terms of sentences of text, in terms of functions, classes, and variables. Yet when we actually want to perform that edit, we have to ultimately tell our editor something of the sort of "delete this character", "replace the next 5 characters by this", ... That's where motions and text objects enter the game. To extend the example above, vim understands w as a motion forward to the beginning of the next word, so 4dw will delete from the current cursor position up to the fourth word to the right.
Text objects can be thought of in a similar way, but without the movement. For example aw is the "outer word" text object, which means "the word under the cursor and the surrounding spaces" (iw, the inner word is without surrounding spaces). Although, I can't move by a text object, I can apply an action such as "delete" to a text object. For example, I can say 4daw to apply (4) times the (d)elete action to the outer word (aw). Similarly, 4dis deletes four sentences (specifically "inner sentences", leaving trailing spaces).
There are text objects for stuff in single quotes, double quotes, parentheses, brackets, braces, function bodies, class definitions, html-tags, variable names, ... Using motions to move around and text objects to specify the target of an action skips that additional step of translating intention into character-by-character operations. And I believe that is what sets vim apart.