Carbon.

[Rodney Sparapani]

Aidan Kehoe wrote:
> 
> It’s not practical though, since for most keys option + key generates a
> distinct key versus the key on its own in the standard Mac maps. (Things like
> œ ˙ é  ® þ ¥ for q w e r t z on the German layout.) And there’s no good way
> to draw a distinction between option + 4, which generates € and is necessary
> for the user, and option + f, which generates ƒ and really isn’t.
> 
> The Aquamacs people approach this in a very ad-hoc way, with per-language
> tables--see
> http://google.com/codesearch?hl=en&lr=&q=file%3Aemulate-mac-keyboard-mode%5C.el%24
> 
> I suppose though a rule of thumb would be ‘if the character generated with option both:
> 
> 1. Differs from the character generated without option and
> 2. Is either ASCII (because our users are programmers) or is necessary to
> write the currently active language
> 
> then option + char-without-option should generate char-with-option;
> otherwise option + char-without-option should generate
> M-char-without-option.’ The information on the characters necessary to write
> a language is available in the language-specific Quail tables, though it is
> incomplete right now. 
> 
> This is horrible. 
> 

I've just starting using Emacs.app for OS X.  And, it works the way I 
would expect.  You can set whichever key that you want to for Meta 
(default alt/option) and still use the Apple/Command key (recognized as 
super).  And, it allows you to do Mule without the problems of Aquamacs. 
  Instead of Carbon, it uses Cocoa/OpenStep/NextStep/whatever you want 
to call it.  Maybe this would be a better basis for the XEmacs port to 
OS X.  I know, patches welcome...  But, if anyone is interested, it can 
be found at http://emacs-app.sourceforge.net/

Right now, everything seems to work except 2-finger scrolling.

Rodney