[ESS] How to avoid garbled characters in output with noweb-mode

[Susumu Tanimura]

Dear Richard, Martin, Douglas Bates, and Douglas,

Thank you very much for suggestions. I did not know useFancyQuotes nor
show.signif.stars opetions in R. I have not used
\usepackage[utf8]{inputenc} so far because I wonder the compatibility
with Japanese pTeX system. A quick trail of
\usepackage[utf8]{inputenc} in euc-jp encoded Rnw source file did not
give me an error. But I may need to check in detail.

> options(show.signif.codes = FALSE)
I guess s/codes/stars/.

--
Susumu

> On Sat, Jan 31, 2009 at 3:28 AM, Martin Maechler
> <maechler at stat.math.ethz.ch> wrote:
> >>>>>> "RMH" == Richard M Heiberger <rmh at temple.edu>
> >>>>>>     on Fri, 30 Jan 2009 23:05:18 -0500 writes:
> >
> >    RMH> I think the easiest way is use the R option
> >    RMH> options(useFancyQuotes=FALSE)
> >
> >    RMH> to tell R to use "'" and not "'" and "'"
> >
> > yes (where the "not .. and .. " really should contain Unicode (UTF-8)
> > backquotes and forward quotes.
> >
> > An alternative that I have not tested for this specific case,
> > but we have been using when working with German umlauts, French
> > accents, etc is to allow LaTeX to work automatically with
> > (many/most) unicode characters *in*  the *.tex file :
> >
> > \usepackage[utf8]{inputenc}
> >
> > in the preamble of your *.tex / *.Rnw document.
> 
> Yet another possibility is to specify in the R
> 
> options(show.signif.codes = FALSE)
> 
> which some of us do all the time because we find those "stars" to be
> objectionable.  Those who want stars on their p-values should be
> required to buy SPSS. :-)