[R] set pdf.options() encoding to UTF-8

[David Winsemius]

On Aug 26, 2009, at 10:55 PM, Liviu Andronic wrote:

> Dear all
> Can anyone point to a list of valid pdf.options() encodings? I checked
> ?pdf and ?postscript, but they do not quite answer my questions. I
> would like to try UTF-8 instead of the default "ISOLatin1.enc" for
> Sweave plots (I have issues with Greek characters in labels).

It's undoubtedly installation specific. The pdf.option page refers us  
(at least on my machine) to pdf() which in turn refers to postscript()  
which says:

encoding
the name of an encoding file. Defaults to "default". The latter is  
interpreted as ‘"ISOLatin1.enc"’ unless the locale is recognized as  
corresponding to a language using ISO 8859-{5,7,13,15} or KOI8-{R,U}.  
The file is looked for in the ‘enc’ directory of package grDevices if  
the path does not contain a path separator. An extension ".enc" can be  
omitted.

So I would be looking where it tells me to look:

/Library/Frameworks/R.framework/Versions/2.9/Resources/library/ 
grDevices/enc

And I see four ISOLatin variants, including ISOLatin2.enc

Obviously that could be different on your machine. I suspect you will  
then need to look up details on the list you find. I seem to remember  
reading that one of the CP125? encodings was similar to UTF-8 but that  
is a very dim memory.

I think that memory might have been regarding this Ripley posting from  
2007:
http://finzi.psych.upenn.edu/R/Rhelp02/archive/91924.html

(... and my memory wasn't very accurate.)

My reading of the R-News article suggest that ISOLatin2.enc might be  
an option for UTF-8 locales:
http://cran.r-project.org/doc/Rnews/Rnews_2006-2.pdf

-- 

David Winsemius, MD
Heritage Laboratories
West Hartford, CT