Re[2]: ill-advised changes [was key bindings in hideshow.el]

[Eric M. Ludlam]

>>> Richard Stallman <address@hidden> seems to think that:
>    It seems that M-o, M-p and M-n are still free.  Maybe one of those
>    could be reserved for minor modes, too?
>
>Maybe it might be reasonable to reserve M-o for minor modes.  M-o is
>not used by Dired as far as I can see.

Not my coveted M-o!  ;)  (I happen to have a C-o like binding to M-o
which I love dearly)


On the topic of Keybindings, and standards:

I've been teaching basic Emacs classes for about 6 years at the
different companies I've worked at.  One thing I always try to explain
in this class is that the keybindings are logical.  "f" means
forward.  Ctrl, Meta, and Ctrl-Meta then do logical things from there.

I think choosing those Meta keys because they are unbound for minor
modes is unwise because the existing C- versions are extensible within
the context of the originating key.  Perhaps a good function for M-o
and C-M-o based on C-o's meaning of opening up a line will present
itself.  For example, my version of M-o is duplicate the current line.
C-M-o as duplicate line, and increment numbers.  They were posted to
ges several years ago and I can re-propose them here if desired.

Based on convention, however, I do think that M-p, and M-n could be
documented as usable by major modes dealing with structured data,
history, or whatever.

Back to Minor Mode standards:

I think there are several kinds of minor modes:

1) Add keyboard-hidden feature (eg auto-fill-mode & eldoc, font-lock)
2) Add complex new functionality and command set.  (eg checkdoc, twocolumn)
3) Folding modes (eg outline-minor-mode, hideshow, folding, etc)
4) Keyboard Emulation (eg tpu, crisp, etc)

I grepped around, and it seems to me that folding modes are special in
that:

a) There are several of them
b) Keybindings should be easy to type to aid navigation.

It may be worth looking at outlining (hiding,showing) as a special
support case for Emacs bindings, and possibly standardizing the tools.

Since outline-minor-mode is well established, it's style of
keybindings could be used.  Of course, there are several C-x postfixes
that aren't being used like c and p.

There are also well used prefixes like C-x r, and C-x v with lots of
extra room in them.

Lastly, after a home for the keys are found, outlining modes could be
granted dispensation for use of extended keys, like [C-End] or [F8] or
whatever for faster typing on X consoles or on Windows.

Once that is worked out, a few base functions, similar to "C-x $"
could be applied, which these minor modes override, providing a
consistent interface to selective display.  Maybe "C-x $" could be
used as one of the controlling keys in this new model.

Back to M-n, and M-p: These keys are already used for structured
buffers for navigation over special types of data.  Since folding
implies structure, these might be useful navigators for a unified
folding keybinding set.  Their use in major modes would then need to
be documented more clearly as to when they can be used.

Anyway, enough mumbling for me.  I don't even use folding modes. :)
Eric