Re: a ergonomic emacs keyboard shortcut layout

On Aug 13, 2007, at 1:39 AM, Peter Dyballa wrote:
> I think your Elisp code is not correctly counting the "self-insert"  
> commands when I enter
>
>         M-x som TAB comm TAB
>
> or do work in *shell* buffer with UNIX commands or type y(es) or n 
> (o) to answer questions. Commands like exit-minibuffer or newline  
> or nil seem to be formal ones I don't ever type. The relation of  
> text input to commands of 1:2 looks strange to me ...

there's nothing much to the elisp program that counts commands.
http://xahlee.org/emacs/command-frequency.el
(it's not written by me, btw)

the key to this short program is just:
(add-hook 'post-command-hook 'command-frequency-record)

the command-frequency-record simply add the value of the buildin var  
“this-command” to a keyed list.

The only thing i found odd is that it lumps commands backward-kill- 
word, kill-word, kill-line, kill-region, all into just kill-region.

it turns out, this is intentional of emacs's “this-command”.  
(thanks to Katsumi Yamaoka).
Quote from lisp doc:

«
— Variable: this-command
    This variable records the name of the command now being executed  
by the editor command loop.
    ...
    Some commands set this variable during their execution, as a  
flag for whatever command runs next. In particular, the functions for  
killing text set this-command to kill-region so that any kill  
commands immediately following will know to append the killed text to  
the previous kill.
»

See the relevant docs here
http://xahlee.org/elisp/Command-Overview.html
http://xahlee.org/elisp/Command-Loop-Info.html

If anyone can contribute a patch to command-frequency.el that'd be  
super!

you wrote:
«...The relation of text input to commands of 1:2 looks strange to  
me ...»

i wondered about that too. But it appears to be a fact. That is, if  
you count cursor moving commands (such as pressing arrow key), one  
actually do that a lot than typing.

See this report:
http://xahlee.org/emacs/ergonomic_emacs_keybinding.html
middle, quote:

-----------------
Better Shortcut Layout vs Better Key Layout

On the subject of keyboarding ergonomics, a user may wonder whether  
switching from QWERTY to Dvorak provides a better improvement than  
switching from a lousy to better keyboard shortcut layout. (Assuming  
that he does only one of the above.)

Of my Emacs Commands Frequency study, 37% of key strokes involves  
data entry (that is, calling the commands “self-insert-command”  
and “newline” (pressing return), while the rest 63% are calling  
all other commands. This seems counter-intuitive, because one might  
think typing should probably be the bulk of activity and moving  
cursor and deleting text or other commands are only done few times  
per minute.

If we include commands invoked by the Backspace key as data entry,  
then 40% are data entry, while 60% are editing commands. Still  
surprising. Going one step further, if we consider the next group of  
most used commands of moving the cursor by the smallest unit (i.e.  
previous-line (C-p), next-line (C-n), backward-char (C-b), forward- 
char (C-f) (also invoked by the physical arrow keys)) as part of data  
entry, then, 64% are data entry and 36% are other command calls.

This data seems to indicate, that a better shortcut layout is as  
important as a better key layout, at least for computer programing  
related text-editing tasks done in emacs.

-----------------

  Xah
  xah@xxxxxxxxxx
∑ http://xahlee.org/



Am 07.08.2007 um 23:49 schrieb xah lee:

> Recently i've done some studies on emacs command frequency and come up
> with a design of ergonomic keyboard shortcut layout for emacs.

I think your Elisp code is not correctly counting the "self-insert"  
commands when I enter

        M-x som TAB comm TAB

or do work in *shell* buffer with UNIX commands or type y(es) or n(o)  
to answer questions. Commands like exit-minibuffer or newline or nil  
seem to be formal ones I don't ever type. The relation of text input  
to commands of 1:2 looks strange to me ...

--
GReetings

  Pete

If all else fails read the instructions.
                                         - Donald Knuth


---------------------------- Info -----------------------------
List Post: <mailto:macosx-emacs@xxxxxxxxxxxxxxxxx>
List Archives: <http://dir.gmane.org/gmane.emacs.macintosh.osx>


☄


---------------------------- Info -----------------------------
List Post: <mailto:macosx-emacs@xxxxxxxxxxxxxxxxx>
List Archives: <http://dir.gmane.org/gmane.emacs.macintosh.osx>