Re: Emacs learning curve

[David Kastrup <dak@gnu.org>]

Richard Stallman <rms@gnu.org> writes:

>     * A scroll bar typically is associated with the view, not the
>     paper: when you 
>     drag it down, the view moves down (and the paper up).
>
>     * A hand pointer typically is associated with the paper, not the
>     view; when you
>     drag it down, the paper moves down (and the view up).
>
> When these command names were chosen, Emacs had neither of those, and
> both analyses made sense. I chose to identify them with the paper.

> Since Emacs has scroll bars, and no hand pointers, it would be more
> consistent today if the scroll commands were named identifying with
> the view.

No, it wouldn't.  You can't exchange the "move in text" view with the
"move the canvas" view interchangeably because Emacs definitely moves
the canvas.  Any point movement associated with it happens as a side
effect of keeping point on-screen.  So if for example you put point in
the middle of your window, then use C-u 3 next (bound to
scroll-up-command), then the canvas will move 3 lines up, but point will
stay put at its buffer position.  The only perceived motion here is
indeed up.

Without the equivalent of some scroll-in-place functionality, the scroll
commands are _not_ a reliable way of moving point through the buffer.
You could claim they are a reliable way of moving window-start through
the buffer.  But I don't think that a user really thinks of window-start
as a separate user-controlled entity.

-- 
David Kastrup