[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: R packages in Debian Science (Was: ITP: r-cran-colorspace -- GNU R Color Space Manipulation)



Belatedly....
On Wed, 2009-01-28 at 09:28 +0100, Andreas Tille wrote:
> On Tue, 27 Jan 2009, Ross Boylan wrote:
> 
> > I haven't been following closely enough to know the purpose, but
> > potentially every R package is (or could be imagined to be) is
> > statistical.
> 
> Sure.  But for instance I'm rather busy packaging r-surveillance
> and other R tools for use in epidemiology.  These will go to a
> med-epidemiology package and do staticts under this topic.  They
> would be definitely missplaced in a science-statistics metapackage.
> 
> The other r-cran-* packages I ITPed are just preconditions and
> while they perfectly fit in the science scope I see better chances
> for categorisation than just trowing them all into the statistics
> bin.  For instance:
> 
>    #512431: r-cran-sp -- GNU R classes and methods for spatial data
>      should go to science geography
> 
>    #512069: r-cran-plotrix -- GNU R package providing various plotting functions
>      (it is ITPed as plotrix but should be renamed to r-cran-plotrix
>       if only ftpmaster whould tell me which way to fix this makes
>       them less work - currently they decided to stay silent at all
>       makes a minimum work for them :-()
>      I tend to put this into science-viewing because it is rather
>      related to creating nice diagrams than doing general statistics.
> 
> So I just wonder whether you see better categories than just
> "statistics" if you see the list of tasks at
> 
>     http://cdd.alioth.debian.org/science/tasks/

I'm not sure if the question is whether the statistics category itself
could be split up, or whether r-cran-msm would be appropriate elsewhere.

Assuming the latter question was intended, statistics seems like a good
place for it.  I believe it came out of epidemiology, and the uses we
are making of it are epidemiological, under a rather expansive
definition of the term.

The statistical model in msm has many potential uses outside of
medicine; it is definitely not just for epidemiology.

I think what that all means is that r-cran-msm would probably also be
good to add to debian-med, as well as having it in science-statistics.

Ross


Reply to: