Re: GNU Emacs raison d'etre

[Karl Fogel]

On 28 May 2020, T.V Raman wrote:
>emacs kbd commands -- and other well-designed ergonomic systems, eg
>vi's h,j,k,l for navigation are better thought of as muscle
>memory. The mnemonics are useful to learn, yes, but given the weird
>layout of the qwerty keyboard,  rigidly sticking to mnemonics often
>leads to non-ergonomic keybindings.

Amen to what T.V. says here.

Often, when people say that keybindings should be "intuitive", they mean 
something like "there should be some connection between a plausible 
English-language description of what the keybinding does and the letters 
involved in the keybinding itself".

But such language/key associations are only useful to newcomers anyway.  After 
all, there is nothing about the word "quit" that inherently suggests its 
meaning -- it's just that those who have learned English have learned what that 
word means.  Similarly, those who have learned the language of Emacs know that 
C-g means the same thing (well, something very similar).

Even independently of keyboard layout (mine is not QWERTY) this kind of 
intuitiveness is of questionable value.  It *does* help newcomers somewhat, but 
if used as an overriding principle it can result in an overly sparse keybinding 
space or in problematic physical combinations like single-finger hurdles.

>So it's always a choice --- does one wish to create a system that is
>"easy to learn" but painful to use, or one that "a little harder to
>learn" with the benefit of being extremely efficient in the
>long-run. I still think VI's nav keys are one of the best choices I've
>seen from an ergonomics point of view, but completely "unintuitive"
>for whatever "intuitive" means.

Agreed.  Vi's default navigation keybindings are, frankly, better than Emacs' 
(or at least they are on a QWERTY keyboard).  It also takes people a long time 
to learn them.

(I'm not suggesting Emacs change its default here: too many people have learned 
the existing way, the efficiency gain is not so huge anyway, and other bits of 
Emacs have been built around the assumptions of those default navigational 
keybindings so there's no telling what full effects of such a switch would be 
at this point.)

Best regards,
-Karl