[R-SIG-Mac] question: cex inside plot()

[Prof Brian Ripley]

On Mon, 23 Jun 2008, Carl Witthoft wrote:

> According to the documentation, the scaling parameter 'cex' is callable 
> inside plot().  However,  I can't get it to change anything.

Which documentation?  It is not mentioned on ?plot.  ?plot.default says

      'cex' a numerical vector giving the amount by which plotting text
           and symbols should be scaled relative to the default.  This
           works as a multiple of 'par("cex")'. 'NULL' and 'NA' are
           equivalent to '1.0'.

so note that this is an argument (not a par), a vector not a scalar and of 
limited scope ('plotting' does not include annotation).

Not that not all plot() methods call plot.defualt(), so there is no 
statement that I can see that all plot() calls will make use of an in-line 
'cex' argument.

> If I set cex using par(),  the graph parameters scale appropriately.
>
> Is this a bug, is it a Mac-only (quartz) bug, or something else?

We need a reproducible example of what does not work.  It is very unlikely 
to be a quartz() bug, as all this is controlled in the base graphics 
subsystem.   Before calling 'bug', please have the courtesy to lay out the 
evidence carefully.

-- 
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