Hmm, I would have recommended
colorRampPalette(c('dark red','white','dark blue'),
space = "Lab")
where the 'space = "Lab"' part also makes sure that a
"perceptually-based" space rather than RGB is used.
I think the functions colorRamp() and (even more)
colorRampPalette() are very nice, part of "standard R" and
still not known and used enough.
Well I use them in ggplot2 :) Unfortunately I've noticed that space =
"Lab" is much slower than space = "RGB" which is why I don't use Lab
space as a default. And it also ignores alpha values, which is a bit
of a pain.
Note that they are based on 'convertColor()' and other color
space functionality in R all of which deserve more usage
in my oppinion and also in my own code ! ;-)
I find colour manipulation generally painful - all the really hard
stuff is there (i.e. conversion between colour spaces), but convenient
functions are lacking. For example, I have this alpha function in
ggplot:
alpha <- function(colour, alpha) {
col <- col2rgb(colour, TRUE) / 255
col[4, ] <- rep(alpha, length(colour))
new_col <- rgb(col[1,], col[2,], col[3,], col[4,])
new_col[is.na(colour)] <- NA
new_col
}
which seems like a lot of work for a simple task. The fact that
col2rgb and rgb aren't symmetric is frustrating, especially since one
outputs values between 0 and 255 and the other (by default) accepts
inputs between 0 and 1.
Package 'vcd' (and others) use package 'colorspace',
and I have wondered in the past if these color space computations
should not be merged into to standard R (package 'grDevices').
But that's really a topic for another thread, on R-devel, not R-help..
It would definitely be nice if all colour space manipulations were
merged into a single package, with a consistent interface. I would
happily contribute to such a project, since I do a lot of colour
manipulation in ggplot.
Hadley