[R] big ps-files

[ripley@stats.ox.ac.uk]

On Thu, 6 Feb 2003, Jonathan Baron wrote:

> On 02/06/03 09:39, Helgi Tomasson wrote:
> 
> >The reason for my question is that a pdf-publication site is asking for a
> >total size of document
> >less than 2MB.   My R-graphs are simple plots of many points. I migth try to
> >tell the same story by
> >using fewer points.  An old (10 years +) Splus gave a particular graph 68k
> >wheras the corresponding
> >R was something like 1MB+.  My latex documents therefore become huge.  The
> >explanation I got was
> >that Splus used some kind of compression optimization which was absent in R.
> >
> >Perhaps it is better to try to generate other form of the graph than ps or
> >to generate the
> >graph using a subsample of the points.
> 
> You say a "pdf-publication."  Pdf itself is compressed.  So when
> you convert to pdf, the figure should get smaller.  Apparently
> this is not happening.

PDF *may* be compressed.  Distilling R ps output makes much smaller 
figures.

> You might try producing pdf directly from R instead of ps.

R's pdf is not compressed, BTW.

> Or use ps2pdf on the output of R, then use pdflatex.

Much better to use Distiller if you have it.

> Or use R's xfig output and then use xfig to produce the eps (and
> then convert to ps), but I'm not sure that xfig will be any
> smaller.
> 
> Or make the graphs smaller using par().  Then, when you include
> them in your LaTeX document, blow them up again using (e.g.) 
> includegraphicx[width=4in]{...}.

That should make almost no difference.

More than 2-3x difference is unusual, but then a 1Mb+ plot from R is 
unusual.  I do suggest you try to make a better graph  as the first step.

We know R's .ps and .pdf output could be optimized and compressed.  But 
that's not something the R developers are interested in, so any 
volunteers?

As a final point, if S-PLUS does what you want, why not use it?  I 
normally do for my publication-quality graphs ....

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595