[Rd] pch=NA (PR#7737)

[Prof Brian Ripley]

Please check the current sources (R-2.1.0 alpha).
pch = NA is now documented under points(), where all the other values are,
and pch="" is now equivalent.


On Tue, 22 Mar 2005 arnima at u.washington.edu wrote:

> I'd like to suggest changes to three help pages, regarding the use of
> pch=NA to suppress plotting symbols. See below.
>
> Arni
>
> R 2.0.1 on WinXP
>
>
>
> ===
> help(bxp)
> ===
>
> The argument outpch=" " needs to be replaced with outpch=NA in two places.
>
> I actually wrote this part of the documentation myself at one point, but
> have now realized that pch=NA and pch=" " are not the same:
>
>   x <- split(rlnorm(26e4), letters)
>   ## NA generates small file
>   postscript("na.ps");    boxplot(x,outpch=NA);  dev.off()
>   ## Space generates large file
>   postscript("space.ps"); boxplot(x,outpch=" "); dev.off()

space was I am sure intended here.

>
>
>
> ===
> help(par)
> help(points)
> ===
>
> Please specify that pch=NA can be used to suppress plotting symbols.
>
> I'm aware of type="n", but pch=NA can be a useful feature inside functions
> and should be documented somewhere. Other negative-sounding values
> (-1,0,NULL,""," ") generate different results, and only pch=NA does the
> job properly.

-- 
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