[R] The steps of building library in R 2.1.1

Sean O'Riordain sean.oriordain at gmail.com
Thu Jul 21 21:24:28 CEST 2005

Just a few thoughts... Good documentation helps everybody - the
beginners and the experts (less beginner questions if there is
thorough and accessible documentation.  I fully appreciate that this
is a volunteer effort - I'm just trying to pin down some places where
we have documentation issues.

Docs can be in a number of different forms - reference, examples,
carefully and throughly explained.  I personally find it difficult to
understand reference type material until I have seen a worked example,
and some of the reference material is a little light on the examples
for me, and others like me who thrive on examples.  The Linux-HOWTO
collection are a good example of step-by-step documentation... If
you're an expert, then you can read it fast and skip... otherwise you
read every line.

Since R comes as a computer "package" and a statistics "thingy"... the
questions on this list come in three forms - those that are very
'package', e.g. "how do I reduce the space between two graphs" - to
the statistics questions "how reliable is the coefficient of
determination in the presence of outliers" (which shouldn't really be
asked here), and then the how do I do 'statistic X' in R - "how do I
calculate a confidence interval around the coefficent of determination
in R?".

The standard documentation got me so far in learning about R - I got a
copy of MASS, S-Programming, "Introductory Statistics with R", and
Michael Crawley's new book "Statistics : An Introduction using R" -
along with every online book I could find on CRAN and elsewhere. 
Unfortunately for me, my experience leaves me in between the beginner
books and the more advanced texts like MASS, David, Schervish etc...
The learning curve is steep - but then like many people, I'd like to
be able to do sophisticated modelling with deep understanding and no
effort :-)

I feel there is a bit of a hole in the middle of the documentation
which could be attacked from both sides - the introduction element is
starting to be covered - it's the next step up from that.

And yes before you ask I would like to help - but my statistics
knowledge is very poor!  Should this conversation go to r-devel?

Thanks for listening,

On 21/07/05, Uwe Ligges <ligges at statistik.uni-dortmund.de> wrote:
> Duncan Murdoch wrote:
> > On 7/21/2005 10:29 AM, Uwe Ligges wrote:
> >
> >>Gabor Grothendieck wrote:
> >>
> >>
> >>>I think you have been using R too long.  Something like
> >>>this is very much needed.  There are two problems:
> >>>
> >>>1. the process itself is too complex (need to get rid of perl,
> >>>    integrate package development tools with package installation
> >>>   procedure [it should be as easy as downloading a package],
> >>>   remove necessity to set or modify any environment variables
> >>>    including the path variables).
> >>>
> >>>2. there is too much material to absorb just to create a package.
> >>>    The manuals are insufficient.
> >>>
> >>>A step-by-step simplification is very much needed.  Its no
> >>>coincidence that there are a number of such descriptions on
> >>>the net (google for 'making creating R package') since I would
> >>>guess that just about everyone has significant problems in creating
> >>>their first package on Windows.
> >>
> >>OK, if people really think this is required, I will sit down on a clean
> >>Windows XP machine, do the setup, and write it down for the next R Help
> >>Desk in R News -- something like "Creating my first R package under
> >>Windows"?
> >>
> >>If anybody else is willing to contribute and can write something up in a
> >>manner that is *not* confusing or misleading (none of the other material
> >>spread over the web satisfies this requirement, AFAICS), she/he is
> >>invited to contribute, of course.
> >>
> >>BTW, everybody else is invited to submit proposals for R Help Desk!!!
> >
> >
> > That sounds great.  Could you also take notes as you go about specific
> > problems in the writeup in the R-admin manual, so it can be improved for
> > the next release?
> Of course.
> > Another thing you could do which would be valuable:  get a student or
> > someone else who is reasonably computer literate, but unfamiliar with R
> > details, to do this while you sit watching and recording their mistakes.
> Good idea.
> Uwe
> > Duncan Murdoch
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html

More information about the R-help mailing list