R-beta: feedback R-0.62.2

Peter Dalgaard BSA p.dalgaard at biostat.ku.dk
Wed Jul 29 14:38:53 CEST 1998

Rashid Nassar <rnassar at acpub.duke.edu> writes:

> Some  observations as feedback on minor points that I came across
> while enjoying [R-0.62.2]:

Thanks for taking the time.

> 1. par(bty="n"): sometimes (depending on data) points appear to
> "spill out" of the plot using default xlim and ylim values (are default
> axes a bit too short?)

As in plot(0.5:4.5,bty="n") ?

This, I believe, is a feature. The axes always extend from the
smallest to the largest *marking*, not the entire x/ylim range. To
get the other effect, use bty="l".

> 2. points(x,y,pch="c") allows points to plot outside existing plot
> box (an interesting feature, but may be a bug?); e.g.,
> > plot(runif(500),runif(500),pch=".")
> > points(rnorm(500)+.5,rnorm(500)+.5,pch="x") 
>                                   #points all over window (pch="c")
> > points(rnorm(500)+.5,rnorm(500)+.5,pch=4,col=2) 
>                                   #points only within box (pch=n)

That one is clearly a bug (still unfixed).

> 3. Typo in locfit() help screen (0.62.1):
>  Value:
>         An object with class "locfit". A standard set of meth-
>         ods for printing, ploting, etc. these objects is pro-
>                           ^^^^^^^

Needs fixing, but this is in a contributed package which belongs to
Clive Loader. If he wants to use alternative speling...

> 4.  Boxplot: The first three work; the last doesn't:
> > plot(A~y)    # A is factor
> > plot(y,A)
> > boxplot(A~y)
> > boxplot(y,A)
> Error: "+" not meaningful for factors

(I think you mean y~A, resp. A,y there.)

This looks strange. It comes from running boxplot.stats on the
grouping variable(s) as well as the analysis variable. I don't get the
point of doing that... For a workaround, use as.numeric(A), but this
seems to require a deeper fix.

> 5. summary(aov.obj) does not print the formula (does not parallel
> output of, e.g., summary(lm.obj) )

That would probably be a good idea.

> 6. using poly(x,n) as argument of lm(): doesn't deal with missing values
> in x; gives error message: 
> Error: NAs in foreign function call (arg 1)

A bug. (Happens in the qr() call, needs some massaging to pass
x[!is.na(x)] and place the results in the right rows of the return
value while NA'ing the others.)

   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

More information about the R-help mailing list