[R] Wikis etc.

David Forrest drf5n at maplepark.com
Mon Jan 9 18:54:30 CET 2006

On Mon, 9 Jan 2006, Gabor Grothendieck wrote:

> On 1/9/06, Thomas Lumley <tlumley at u.washington.edu> wrote:
> > On Mon, 9 Jan 2006, Michael Dewey wrote:
> > >
> > > Further to that I feel that (perhaps because they do not like to blow their
> > > own trumpet too much) the authors of books on R do not stress how much most
> > > questioners could gain by buying and reading at least one of the many books
> > > on R. When I started I found the free documents useful but I made most
> > > progress when I bought MASS. I do realise that liking books is a bit last
> > > millennium.
> > >
> >
> > Very late last millenium, though.
> > "When I were young[er] we didn't have all these fancy yellow books."
> >
> > More seriously, yes, reading books about R and S is very effective and is
> > how most of the R experts learned.  In my case it was the Blue Book, the
> > White Book, and the Ripley/Venables/Smith notes on S-plus (which have
> > evolved to the Introduction to R).
> In addition to books, the various manuals, contributed documents and
> mailing list archives, all of which one should review,
> the key thing to do if you want to really learn R is to read source code
> and lots of it.  I think there is no other way.  Furthermore, the fact that
> you can do this is really a key advantage of open source.

There has to be some reason to dig into the source code.  Just starting at
line 1 and reading until you are enlightened would be frustrating,
repetetive, and nearly pointless.  The great benefits of the R books is
that they have interesting results (a fancy graph, analysis, or report)
that you can trace back to the constituent parts and (with open source
code) learn everything you want to and be confident that the rest is there
if you need it.  Books using R do an excellent job of showing what is
possible and, through recursive study of the open source code, how to do
it.  Books connect high-level tasks to low-level functions.

R has plenty of documentation, but if the measure of excellence is simply
number of pages or weight, SAS's documentation might still win even if we
include the pages of R source code.  Both packages have lots of detailed
documentation, where if you understood everything that was written, you'd
know how to do what you want.  The authors of neither R nor SAS have
failed to document their functions.

Where I think the R (and SAS) documentation is lacking is in the
connections between the documentation elements.  RTFM isn't helpful if you
can't find TFM.  For instance, there is more than one way to make a graph,
(see http://addictedtor.free.fr/graphiques/thumbs.php?sort=package for 135
of them) how does a novice know which function to use?  How do you find
out the alternate ways to do things?  The hateful MS Excel solves this by
registering the alternate graphic capabilities under a hierarchical GUI
menu.  We solve it with an email list and several fuzzy searches.

Since R has such an extensive set of extensions, maybe we need a section
in the R-intro documentation near
titled "Finding existing functions".  It could explain the difference
between base and recommended, installed, CRAN, and how someone can find
and use things in these areas using help(), '?', help.search(),
help.start(), RSiteSearch(), and the mailing lists.

 Dr. David Forrest
 drf at vims.edu                                    (804)684-7900w
 drf5n at maplepark.com                             (804)642-0662h

More information about the R-help mailing list