#	START NEW ARIX SCCS HEADER
#
#	@(#) README: version 25.1 created on 12/2/91 at 19:08:32
#
#	Copyright (c) 1990 by Arix Corporation
#	All Rights Reserved
#
#	@(#)README	25.1	12/2/91 Copyright (c) 1990 by Arix Corporation
#
#	END NEW ARIX SCCS HEADER
#
#ATT	ident	"curses:demo/pacman/README	1.2"

#ident	"@(#)libcurses/demo/pacman:README	25.1"

Pacman has been tested on cbosgd (running 4.1BSD with the fast timer
driver added), and on vanilla 5.0 systems on a VAX and on a 3B.
It is strongly urged that you add the fast timer driver (../ft) since
otherwise the nap mechanism won't work very well.  (On 5.0, for
example, it will feel strange every time you hit a key.)
To my knowledge, ft needs to be ported to 5.0, and I don't know of
anyone who has done this.

Pacman should work on any system that can run this version of curses.
The worst that will happen is that turns will only go one per second,
resulting in a very slow (and boringly easy) game.  You need some form
of improved nap to make the game go fast enough to be interesting.

Recent changes to Pacman make it work reasonably well at 1200 baud on
verbose terminals such as HP's, vt100's, and Ambassadors.  However,
it still feels better at high speeds and on terse terminals such as
the mime, concept, blit, and adm3a.  ("verbose" and "terse" refer to
the length of the cursor addressing sequence - 3 or 4 characters is
terse, 8 is verbose.)  The improvement is to use the draino function,
which needs TIOCOUTQ, which, alas, only exists on 4.1BSD.  (But all
the ioctl does is return the number of characters in the output queue,
so it ought to be trivial to add to 5.0.)
