[R] Thanks

[Prof Brian Ripley]

On Tue, 7 Sep 2004, Liaw, Andy wrote:

> A small drawback to Gabor's proposal:
> 
> > "aa" %:% "dd"
> [1] "b" "c" "d"
> > "00" %:% "99"
> character(0)
> 
> Defining it as a seq() method for characters is probably not a terribly good
> idea, as one should expect it to work for any character vectors as input.
> (One can argue that the function worked as `expected', I suppose...)

Another drawback: the order of characters is locale-dependent. letters and
LETTERS are an Anglophone's view of what happens (or happened, probably)  
in Roman.

-- 
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