[R] persp or scatterplot3d?

[Richard A. O'Keefe]

heimdal at aracnet.com
	[has three x-y plots with the same horizontal axis and similar
	 vertical axes.  He wants to see the three variables together,
	 but if the graphs are plotted in the same rectangle, they
	 obscure each other.  He wants to make a "visually useful" plot.]

Why not just do
    oldpar <- par()
    par(mfrow=c(3,1))
    plot(x, y1)
    plot(x, y2)
    plot(x, y3)
    par(oldpar)


	y is the value of "io", "sys", or "usr" that occurs at time "x".

This sounds like the amount of time being spent by a computer on
various activities.  A plot that might be revealing, although it
would lose the time dimension, would be a ternary plot (see
help("ternaryplot", package="vcd")); the time dimension could be
partly recovered by chopping the times into equal-length subintervals
and plotting a separate ternary diagram for each, in a matrix of
ternary diagrams.