Aspect ratio and limits
On Tue, 2005-04-19 at 17:40 +0100, Barry Rowlingson wrote:
Marc Schwartz wrote:
# Set height value as you require height <- 3 X11(width = aspect * height, height = height) .... Does that get you what you need?
I'm guessing that will work, but it just seems unsatisfactory, since if I resize my graphics window I mess it up.
Patient: "Doctor, when I move my arm like this it hurts." Doctor: "Well...stop moving your arm like that." ;-) I think that unlike your lattice example below, the X11 device is "elastic" on a resize. I do not see a par() setting that alters that behavior. For some reason, I think that under Windows there was a setting for the default Windows graphics device relative to resizing, but that my be the result of a TIA and not the representation of reality...I hate it when that happens. :-)
If I set par(pty='s') I can have a square plot floating inside a graphics window of any shape, I'd like to be able to plot something specifying ylim and asp and get a rectangular plot with a fixed aspect ratio floating around inside a resizable graphics device.... The 'aspect' ratio of xyplot (lattice package) lets you specify the aspect ratio of the plot (not relating to the data, ie aspect=1 draws a square plot whatever the ratio of X and Y range). Try this: xy=data.frame(x=c(0,21),y=c(0,4)) xyplot(y~x,data=xy,aspect=0.2) - now resize the graphics window and the plot keeps its ratio. That's the effect I was hoping to get from base graphics. I could use lattice graphics but I'd have to work out the value of 'aspect' to get the 1:1 aspect ratio for my data. I'm not sure this is trivial, since it may involve considering the various margins and spacing around a plot. Close...
There is a fair amount of interaction there. I am not sure what might be a "best" approach. I'll think about it. If I come up with anything, I'll let you know. I guess that one underlying consideration I might think about here is, where is your final output going. If to a PS file for printing or inclusion in a LaTeX document, I would focus on how the plot looks there, especially when trying to compare it to X11 output. Marc