RE: [External] : Re: Experimentally unbind M-o on the trunk

[Drew Adams]

> > E.g., I see this if I type just `save' to the
> > prompt from `lacarte-execute-menu-command':
> >
> > Buffers > Frames > SAVE
> > Buffers > SAVE  %
> > Do Re Mi > Save Frame Configuration          (C-x t .)
> > File > Bookmarks > Here (This File/Buffer) >
> >   Save Bookmarks Here To Bookmark File...    (C-x x C-s)
> > File > Bookmarks > Save Bookmarks As...      (C-x x w)
> > File > Bookmarks > Save Bookmarks            (C-x x s)
> > File > Open Recent > Save list now
> > File > Save As...                            (C-x C-w)
> > Icicles > + Remove Saved Candidate Set...
> > Icicles > Add/Update Saved Candidate Set...
> > Icicles > Save String to Variable...
> > Options > Customize Emacs > Saved Options
> > Options > Icicles > Toggle > Highlighting Saved Candidates
> > Options > Save Frame Configs (DoReMi)
> > Options > Save Options
> > Options > Save Place in Files between Sessions
> > Tools > Spell Checking > Save Dictionary
> >
> > (That's with substring completion.  And some of
> > those menus are from my code.)
> 
> How is this different from "M-x save- TAB"?

I don't understand the question.  The context was
being able to efficiently use menus.

It's not about opposing use of menus to use of
command-names.  It's about showing that completion
of menus offers the same benefits as completion of
command names.

This was a response to your statement to the effect
that it's not possible to use menus efficiently,
particularly to be able to access something anywhere
in the entire menu forest (as opposed to within a
particular menu).

You had said, "That only works well if you can guess,
up front, which top-level menu item has the command
you are looking for."

I tried to show that that's not inherently the case.

> > That's a pretty darn quick way to search the entire
> > menu-bar forest, to see all menu items whose paths
> > contain the word "save".
> 
> The goal was not to search the menus, the goal was to discover
> commands.  Menus were suggested as a means towards that end.  Let's
> not confuse the two.

Your statement was in reply to this:

  The discovery scenario is: I don't know what I'm looking for, but I
  can progressively narrow down the command space by choosing a submenu
  at each step. Once I've found the command I need, I can execute it
  right away and be done with it.

I read that with found-the-command meaning just
getting access to the command.  I thought/think
Yuri was saying that menus are a good way to
discover how to do something.

Finding out the name of the command that corresponds
to a given menu item is something else.  I didn't
take that to be what Yuri meant by discovering the
command.

But that too can be helped with the same or similar
tools.

For example, with Icicles you can just hit a key to
get the command name of any of those menu completion
candidates.

In fact, you can see their full doc strings in *Help*,
on demand while menu-completing.  And even without
doing that, just by cycling among candidates you can
see short help about their commands in the mode line
of buffer *Completions*.