[R] charset in graphics

[Prof Brian Ripley]

On Fri, 13 Jul 2007, Peter Dalgaard wrote:

> Prof Brian Ripley wrote:
>> On Fri, 13 Jul 2007, Peter Dalgaard wrote:
>>
>>> The correct incantation seems to be
>>>
>>> postscript(font="URWHelvetica", encoding="ISOLatin7")
>>> plot(0,main=tolower("\u104\u116\u0118\u012e\u0172\u016a\u010c\u0160\u017d"))
>>> dev.off()
>>>
>> The encoding should happen automagically in a Lithuanian UTF-8 locale, and
>> does for me.  But suitable fonts (e.g. URW ones) are needed.
>>
> OK, I sort of suspected that, although it wasn't entirely clear to me
> whether autoconversion would cover cases like en_LT.utf8, if that even
> exists.

The locale would need to be in the Lithuanian language (ISO 639 code lt or 
lit), not some language in Lithuania (ISO 3166 code LT).

> Still, the explicit (portable?) way of doing it is probably
> worth knowing too (there could be a few pitfalls with scripts getting
> run outside their usual domain).

Yes, the appropriate 8-bit encoding can only be a guess, and someone might 
be writing French or Japanese in lt_LT.utf8.

-- 
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