Martin Maechler
2003-Oct-15 23:47 UTC
[Rd] qqnorm(*, datax=TRUE, xlab,ylab) -- S+ compatibility problem
Just found because an old (written for S+) function of mine did
label plots wrongly with R.
Example --- inspired from example(qqnorm) ---
data(precip)
qqnorm(precip, ylab = "Precipitation [in/yr] ...")
qqnorm(precip, ylab = "Precipitation [in/yr] ...", datax = TRUE)
this is all "fine" -- `datax' is a switch that just switches the
axes but keeps the meaning of "ylab labels the data axis"
(i.e, when changing datax from FALSE to TRUE, you don't have to
switch xlab and ylab in your call).
Unfortunately, S+ (6.1) does this differently,
namely with the logic of "ylab labels the y-axis",
i.e. the last line above would have to be written
qqnorm(precip, xlab = "Precipitation [in/yr] ...", datax = TRUE)
## _^_
for S+ to produce the same (well analogous) plot.
Since, the `datax' extension was only introduced for R 1.7.0,
with
NEWS> o qqnorm() and qqline() have an optional logical argument
NEWS> `datax' to transpose the plot (S-PLUS compatibility).
^^^^^^^^^^^^^^^^^^^^
we could consider changing R's behavior here
--- unfortunately not back-compatibly.
Opinions?
Martin Maechler <maechler@stat.math.ethz.ch>
http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum LEO C16 Leonhardstr. 27
ETH (Federal Inst. Technology) 8092 Zurich SWITZERLAND
phone: x-41-1-632-3408 fax: ...-1228 <><
