redirected to xemacs-beta, this needs more discussion.
>>>> "anthony" == anthony 
<anthony(a)magix.com.sg> writes: 
    anthony> As an end-user, I hit page-up or page-down to move
    anthony> through a file. These are bound to
    anthony> scroll-{up,down}-command, which behave much as I expect,
    anthony> except in the end-case when the screen cannot be scrolled
    anthony> any further, when the command fails.
This is definitely the behavior I prefer.  When moving by screen I
prefer point to be approximately in the middle of the window.  If I
wanted end-of-buffer, I'd use M->.  So unless there nearly unanimous
support from users/the Review Board, I would be inclined to veto this
patch.
    anthony> But it's still possible to "move further", since the
    anthony> point is still in the centre of the screen.  So with this
    anthony> patch, point is then moved to the {start,end} of the
    anthony> buffer.  Subsequent attempts to scroll
    anthony> maybe_signal_error_1 as before.
This isn't unreasonable, though.  If it were switchable by a variable
I would be inclined to approve it.  I'd probably push for your
preference to be the default in the beta tree as that's the best way
to get feedback; we can revert to current later if more people hate it
than like it.
Comments?  (Patch omitted, it's on xemacs-patches.)
-- 
Institute of Policy and Planning Sciences     
http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
              Don't ask how you can "do" free software business;
              ask what your business can "do for" free software.