[Expm-developers] R package 'expm' - release to CRAN?

[Martin Maechler]

>>>>> "VG" == Vincent Goulet <Vincent.Goulet at act.ulaval.ca>
>>>>>     on Fri, 18 Jun 2010 07:57:01 -0400 writes:

    VG> Hi Martin,
    VG> Thank you for this initiative.

    VG> I wholeheartedly support releasing expm to CRAN. Any chance it becomes a Recommended package? If Matrix depends on it, that would make sense.

Dear all, I'm sorry, but I've been overwhelmed with all kinds of
activities these days.

Before the useR conference, I'd only like to do 2 things:

- Change the official package maintainer from 
  	 expm-developers at r-forge.wu-wien.ac.at
  to myself.
  I've really done almost all of the recent changes (though with
  very important code contribution from Michael Stadelmann), and have
  been clearly the lead developer..

- release it to CRAN "as is"

-------------------------------------------------


The following has to come later for sure :


    VG> Also, can we get a publication out of this, say in JSS? I'd be willing to contribute my fait share in this area.


Indeed, I'd be very glad if some of my (and your) nice
experiments and examples could make it into a vignette (Sweave:
R & LaTeX together) which we'd submit to JSS.
I don't think that all current expm-developers should be
co-authors though, but it would be great if you Vincent would
take the lead on this.
At the moment, it would seem sensible to think of an author
"trium virate"
  Vincent, Daniel (supervisor of Michael) and me,
of course acknowledging the work of the other expm-developers.

Martin

    VG> Best,

    VG> Vincent

    VG> Le ven. 18 juin à 06:47, Martin Maechler a écrit :

    >> Dear 'expm' co-developers,
    >> 
    >> After Michael Stadelmanns master thesis finished here at ETH,
    >> more than a year ago, which provided one or two new state-of-the-art matrix
    >> exponential algorithms, and the Matrix-Log + Matrix-Sqrt,
    >> not much has happened
    >> though I've noticed that Nick Higham has produced yet another
    >> improvement to his algorithm, namely: 
    >> 
    >> author =    {Awad H. Al-Mohy and Nicholas J. Higham},
    >> year =      2009,
    >> title =     {A New Scaling and Squaring Algorithm for the Matrix Exponential},
    >> journal =   {SIAM Journal on Matrix Analysis and Applications},
    >> volume =    31,
    >> number =    3,
    >> pages =     {970-989},
    >> 
    >> ( I have added the reference to expm/inst/doc/expm.bib )
    >> 
    >> But our 'expm' package is still only on R-forge and hence not
    >> really much visible in the R using comunity.
    >> Also, the recommended Matrix package which has provided expm()
    >> for a long time, still only provides "Ward77"
    >> which is clearly sub-optimal given Higham's work.
    >> 
    >> OTOH, in some way, 'expm' is now an "extension" to the Matrix package
    >> (as it 'Depends' on it anyway, e.g. for using  Matrix::norm()),
    >> providing extra features.
    >> 
    >> In light of this, I'd like to release 'expm' to CRAN
    >> ASAP, and just wanted to get feedback from you before doing
    >> that.
    >> 
    >> Best regards,
    >> Martin
    >> 
    >> Martin <Maechler at stat.math.ethz.ch>  http://stat.ethz.ch/people/maechler
    >> Seminar für Statistik, ETH Zürich  HG G 16      Rämistrasse 101
    >> CH-8092 Zurich, SWITZERLAND
    >> phone: +41-44-632-3408       fax: ...-1228      <><
    >> _______________________________________________
    >> Expm-developers mailing list
    >> Expm-developers at lists.r-forge.r-project.org
    >> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/expm-developers

    VG> _______________________________________________
    VG> Expm-developers mailing list
    VG> Expm-developers at lists.r-forge.r-project.org
    VG> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/expm-developers