[R-sig-Geo] (no subject)

[Roger Bivand]

On Thu, 28 Sep 2006, Frank Hardisty wrote:

> Dear Jacques Vernin (and All),
> 
> I appreciate that you are experiencing some difficulties when using R
> and spatial extensions.
> 
> However, please realize that Roger Bivand has been wonderfully patient
> and helpful to many, many people, as you can see in the history of
> this email list.
> 
> Please remember that R and R spatial extensions are written,
> maintained, and supported by volunteer labor.

Thanks for your kind words, Frank! But in fact at least one of my replies
to Jacques was brusque, I can get annoyed when people do not follow
general R list advice and state their platform and package versions. So
his choice of tone is partly my choice. 

The resolution here is that overloading existing functions is *not a good
idea* and can lead to unexpected consequences, like this method
dispatching muddle. Using the package framework and guarding the local
code in a namespace should solve it.

Roger

PS. I had a very interesting talk at FOSS4G with someone attacking the 
spatial R/Java puzzle through JRI (see the rJava package), who reported 
much more success than we had two years ago.

> 
> Your email "tone of voice" is not very sweet, considering that you are
> asking for free help.
> 
> regards,
> -Frank Hardisty
> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Frank Hardisty
> Department of Geography
> University of South Carolina
> hardisty at sc.edu
> Office -- 803-777-5729
> Fax -- 803-777-4972
> http://people.cas.sc.edu/hardistf/
> 
> 
> On 9/28/06, Frank Hardisty <HardistF at gwm.sc.edu> wrote:
> >
> >
> > ---------- Forwarded message ----------
> > From: "pir2.jv" <pir2.jv at wanadoo.fr>
> > To: Roger.Bivand at nhh.no
> > Date: Thu, 28 Sep 2006 16:42:13 +0200
> > Subject: Re: [R-sig-Geo] (no subject)
> >
> > Le 28 sept. 06 à 16:18, Roger Bivand a écrit :
> >
> > >>>
> > >>> If you say ls(), do the names of any of the objects begin with plot?
> > >>
> > >>> ls(pat="plot")
> > >> [1] "BBplot"             "Bplot"              "draw.plotmath.cell"
> > >> [4] "plot"               "plotG"              "plotM"
> > >> [7] "plotx"              "plotxx"
> > >>
> > >> plotx, plotxx, plotG and plotM are functions of mine. Is there a
> > >> problem ?
> > >
> > > The problem is that your local function plot is masking the one which
> > > understands that "SpatialPolygons" objects have their own method
> > > and need
> > > to be sent there, so yes, at least your local plot is the problem. The
> > > standard advice if you have a number of functions is to write a
> > > package
> > > with a NAMESPACE, and that will protect you from R not choosing the
> > > one
> > > you intended.
> > >
> > Why would be my function plot local ? It is not.
> > I removed the others
> >
> > ============
> >  > plot
> > function (x, y, ...)
> > {
> >      if (is.null(attr(x, "class")) && is.function(x)) {
> >          nms <- names(list(...))
> >          if (missing(y))
> >              y <- {
> >                  if (!"from" %in% nms)
> >                    0
> >                  else if (!"to" %in% nms)
> >                    1
> >                  else if (!"xlim" %in% nms)
> >                    NULL
> >              }
> >          if ("ylab" %in% nms)
> >              plot.function(x, y, ...)
> >          else plot.function(x, y, ylab = paste(deparse(substitute(x)),
> >              "(x)"), ...)
> >      }
> >      else UseMethod("plot")
> > }
> > <environment: namespace:graphics>
> >
> > =============================
> > >
> > >
> >
> > Jacques Vernin
> >
> > _______________________________________________
> > R-sig-Geo mailing list
> > R-sig-Geo at stat.math.ethz.ch
> > https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> >
> >
> 
> 
> 

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no