[R] Problem with Palatino font in pdf figures

[Prof Brian Ripley]

On Sun, 30 Sep 2007, Paul Smith wrote:

> On 9/30/07, Marc Schwartz <marc_schwartz at comcast.net> wrote:
>>> Consider the following piece of code:
>>>
>>> pdf(file="figure.pdf", family="Palatino")
>>> plot(0,0,type='n', xlim=c(-20,20), ylim=c(0,2),xlab="",ylab="",axes=F)
>>> text(-1.4,1.168,expression(italic("The font looks different when this
>>> is seen with Acrobat Reader!")),xpd=T)
>>> dev.off()
>>>
>>> When viewing the produced figure.pdf with kpdf (on Linux), it looks as
>>> being written with LaTeX Mathpazo font, but not when one views
>>> figure.pdf with Acrobat Reader. Any ideas about how to get the same
>>> result both with kpdf and with Acrobat Reader?
>>
>> Note that R does not embed the fonts by default in the PDF file.
>>
>> Based upon what I am seeing here on F7, which is consistent with your
>> comments, the font substitution mapping in Adobe Reader 8 is different
>> than that in either kpdf, gv or in Evince. The latter three appear to be
>> using the same font substitution and look the same.

There is a note to that effect in ?pdf.  All the viewers based on 
GhostScript are going to be using URW fonts (under Linux, at least), that 
is use URW Palladio to substitute Palatino.  xpdf is configurable, but I 
believe is also normally set up to use URW T1 fonts.

>> You might want to review ?pdfFonts and ?embedFonts for additional
>> information as well as the article by Paul Murrell and Prof. Ripley in R
>> News on non-standard fonts:
>>
>>   http://cran.r-project.org/doc/Rnews/Rnews_2006-2.pdf
>
> Thanks, Marc. With embedFonts, everything gets right!

I dobut it: Palatino is a commercial font (I believe it is a Linotype 
trademark) and _if_ you have it you need to worry about the legality of 
embedding it.  Most likely you have arranged to substitute URWPalladio 
everywhere, in which case you would do better to use the correct font 
metrics by specifying that in the first place.

-- 
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