Input methods at the age of Unicode

Richard Wordingham richard.wordingham at
Fri Jul 17 01:39:57 CDT 2015

On Fri, 17 Jul 2015 05:41:11 +0200
"Janusz S. Bien" <jsbien at> wrote:

> Quote/Cytat - Richard Wordingham <richard.wordingham at>  
> (Fri 17 Jul 2015 12:59:24 AM CEST):

Perhaps I'm missing a trick.  My conception was that to use an Emacs
keyboard for, say, word processor input, one would have to type into
an Emacs buffer and then copy the text to the word processor

> > it's easy to extend it own
> > keyboards.  (Creating the first one was a bit stressful
> It is not clear for me what do you mean by "own keyboards"

Except possibly for Windows (last time I looked into it, Emacs there was
built as an ANSI application rather than as a Unicode application),
Emacs can use the user-specified system keyboards (and general-purpose
user keyboards) as well as the Emacs-specific keyboards.  By "own
keyboards" I meant the ones defined for Emacs, specifically the ones
set up by quail-define-package and quail-define-rules.

There was a period when, due to an external error, Emacs launched with
an English locale couldn't use keyboards made available by ibus.

> - the ones
> > that come with Emacs were almost all set up using ISO-2022, before
> > Emacs adopted Unicode.)

> I my opinion creating a new Emacs input method is extremely easy and
> I solve my problems my modifying 'polish-slash'.

I see latin-pre.el and latin-post.el in particular are now defined in
UTF-8, which simplifies adaptation.  My exemplar was thai.el, which
at the time was in ISO-2022.

> In a file you can associate an input method with it using Emacs an  
> appropriate local variable.

Another example of the first keyboard being difficult and the rest
easy.  Once one starts using that trick it is easy to modify it for
other keyboards.


More information about the Unicode mailing list