Re: Customizing key bindings (was: Re: [CVS] f7, f8 bound..)

[Robert J. Chassell]

Per Abrahamsen <address@hidden> wrote:

   The problem is the mindset that making complex applications written
   by programmers for programmers easy to configure, is simply a
   matter of providing a "friendly" interface to the existing
   configuration mechanisms.  This mindset is shared by most
   traditional Unix developers (and probably Lisp developers too, but
   who cares).

   However, this assumes that being "easy to configure" is merely a
   matter of "syntactical sugar", that is, provide a form based
   interface to the old text based configuration files.

Please explain in more detail.  I cannot infer your implications.

To me, the current customize is "syntactical sugar".  That is to say,
it creates, edits, and writes expressions to my .emacs file for two
functions, `custom-set-faces', and `custom-set-variables'.

I can edit and evaluate those function calls in my .emacs, or I can
edit and evaluate them using a `Customize' buffer.  Either way, I end
up with expression in my .emacs file.

   If you want to make configuration easy to the end-users, you need
   to start with their needs, and not with the old textual
   configuration files.

What are the implications of this?

After all, "the old textual configuration files" are not the `surface
representation' that an end-user ever sees; they are a `deep
structure' that a programmer sees, if he or she wishes.  (To the
programmer, of course, they are a `surface representation' of the
underlying bits.)

The configuration *must* be saved somehow.  Otherwise, the
customization did not `take hold'.  Why not save it in a format that a
programmer can readily edit?  (And in the case of Emacs, in a format
that the existing virtual machine can readily evaluate?)

-- 
    Robert J. Chassell            address@hidden  address@hidden
    Rattlesnake Enterprises       http://www.rattlesnake.com
    Free Software Foundation      http://www.gnu.org   GnuPG Key ID: 004B4AC8