R devel - Nov 2008 - Computing minimal detectable differences for general ANOVA models

Dear Developers,

I would like to contribute to the R Project.

I implemented a C function the can be used to calculate minimal
detectable differences for general ANOVA models (sample size
calculations, power analysis).

A demo program and the source code is available from my website:

http://reliablecomputing.eu/ncbeta.html

What should i do in order to make my ncbeta function part of the R
project source distribution?

Probably the most troublesome parts of my code are those where error
may occur, i do not know much about the proper error handling in R.

Many thanks,

Ali

P.S. I see my suggestions
https://stat.ethz.ch/pipermail/r-devel/2008-April/049317.html
were considered in R 2.8.0.

On Sun, 16 Nov 2008, Ali Baharev wrote:
> Dear Developers,
>
> I would like to contribute to the R Project.
>
> I implemented a C function the can be used to calculate minimal
> detectable differences for general ANOVA models (sample size
> calculations, power analysis).
>
> A demo program and the source code is available from my website:
>
> http://reliablecomputing.eu/ncbeta.html
>
> What should i do in order to make my ncbeta function part of the R
> project source distribution?
Everyone can contribute code as a package: code in the source distribution 
is by members of R-core only, including in a few cases code they choose to 
adopt (with permission) from others.  Most importantly, R-core has to 
commit to maintenance of such code indefinitely, and these days we are 
very reluctant to do so.
> Probably the most troublesome parts of my code are those where error
> may occur, i do not know much about the proper error handling in R.
So you need to find out before distributing a package.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

Ali Baharev wrote:
> Dear Developers,
> 
> I would like to contribute to the R Project.
> 
> I implemented a C function the can be used to calculate minimal
> detectable differences for general ANOVA models (sample size
> calculations, power analysis).
> 
> A demo program and the source code is available from my website:
> 
> http://reliablecomputing.eu/ncbeta.html
> 
> What should i do in order to make my ncbeta function part of the R
> project source distribution?
As Brian said in  a separate message, build an R package
and distribute one or more R documented functions that
all the C code.
See the Writing R Extension's manual.
> 
> Probably the most troublesome parts of my code are those where error
> may occur, i do not know much about the proper error handling in R.
Calls to exit() as in your C code should be replaced with calls to

    PROBLEM "the error message"
    ERROR

Again, the R manuals (Writing R Extensions) will give you the
information you need.
> 
> Many thanks,
> 
> Ali
> 
> P.S. I see my suggestions
> https://stat.ethz.ch/pipermail/r-devel/2008-April/049317.html
> were considered in R 2.8.0.
> 
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

Ooops, sorry. Just one more question.

Do i have to make a binary package for Win32 and Mac OS X? Or is it
done by some server side scripts?

For Win32 it seems to me

http://win-builder.r-project.org/

can build it for me.

Sorry for the dumb questions.

Ali

They will be built for you.  You only upload the .tar.gz file.
See section 1.5 of the Writing Extensions manual.

On Sat, Nov 22, 2008 at 5:35 PM, Ali Baharev <ali.baharev at gmail.com>
wrote:
> Ooops, sorry. Just one more question.
>
> Do i have to make a binary package for Win32 and Mac OS X? Or is it
> done by some server side scripts?
>
> For Win32 it seems to me
>
> http://win-builder.r-project.org/
>
> can build it for me.
>
> Sorry for the dumb questions.
>
> Ali
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>