[Rd] yinch() is NOT nonsense with log='y'

[Spencer Graves]

       May I request a modest change in the documentation and warning 
for "xinch", "yinch", and "xyinch"?


       Consider the following example:


 > plot(1:2, log='y')
 > yinch(1)
[1] 0.1961134
Warning message:
In yinch(1) : y log scale:  yinch() is nonsense


       In fact, in this environment, "yinch(y)" translates a vertical 
displacement of "y" inches into the corresponding change in the log10 
scale.  For a realistic example of such a use, see the first example of 
the current R-Forge version of the new "animate" function in the "Ecfun" 
package:  This places the Rlogo on a plot(..., log='y') without distortion.


       Thanks,
       Spencer Graves