R-beta: Plot and scale

George White gwhite at caligo.bio.dfo.ca
Thu Mar 19 20:20:48 CET 1998

On Thu, 19 Mar 1998, Bill Simpson wrote:

> On 19 Mar 1998, Peter Dalgaard BSA wrote:
> > Bill Simpson <wsimpson at uwinnipeg.ca> writes:
> > 
> > > postscript(...,width=..., height=...) command.  I personally think it is a
> > > design error in S-Plus/R for postscript() to modify the appearance of the
> > > plot compared to how it looks in the Graphics window.
> > 
> > Mmm....no. Not unless you want all devices to give identical output.
> > Different devices have different characteristics, media sizes, physical
> > fonts, etc. postscript() and x11() are just two different device
> > drivers. 
> I knew people would disagree.
> I have this opinion from years of working on the Mac, where the WYSIWYG
> philosophy permeated everything.  The monitor and the printer were always
> coordinated so the physical dimensions of a figure on the monitor and the
> physical dimensions of the figure on the printer were identical.  (I
> am including fonts of course) This was achieved despite the huge diffs in
> monitor and laser printer resolution. I think this approach was also used
> on NeXT.  And it is a wonderful time-saver for people who do graphics.
> If you haven't experienced it, you don't know what you're missing.

Very true.  Also, colors on NeXT were well matched between monitor
and output devices.  

> Right now the only way to make figs look the way you want in R (X11) is
> through trial and error (very time-consuming). (BTW using Mac R gives
> figures with the lovely monitor=printer characteristics I've been talking
> about, due to system software)
Well, you can take a couple measurements and deterine the pixel size
(and choose a mode that gives square pixels!).  Color is a much harder
problem unless you are willing to limit yourself to a small set of

> I know this lack of output device coordination was a design decision
> (maybe not by S-Plus but at level of X-windows etc).  But it was a bad
> decision.

Workstations and X-terminals generally have known pixel sizes (usually 72
or 100 dppi) and can be used to get consistent sizes (and have good color
control).  Using XFree86 I do miss the lack of control over these
factors, but it does give a lot of bang for the $ when I'm not fussing
over publication quality output. 

I do hope that these issues will eventually be addressed in XFree86.

George White <whiteg at mar.dfo-mpo.gc.ca> <aa056 at chebucto.ns.ca>  902.426.8509 

r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

More information about the R-help mailing list