R help - Jun 2011 - Standards for delivery of GPL software in CRAN packages

I wondered if there were standard practices in CRAN for delivery of R source
implementing functions in R packages. I has encountered a couple of packages
where the gzipped version of source contains very little, primarily the Help
files describing the functions in the package. In some cases I can find the
source as the value of the function name.

Given that these packages are released as GPL, oughtn't the unoptimized
source be freely available, hopefully with comments? Am I missing something? Is
there a central place other than mirrors where such source is retained?
Sourceforge?

  - Jan

- Jan, from Sierpinski, 
     a Blackberry, 6072391834,
     Google Talk to: 
       bayesianlogic.1 at gmail.com

On Mon, Jun 27, 2011 at 1:24 PM, Galkowski, Jan <jgalkows@akamai.com>
wrote:
> I wondered if there were standard practices in CRAN for delivery of R
> source implementing functions in R packages. I has encountered a couple of
> packages where the gzipped version of source contains very little,
primarily
> the Help files describing the functions in the package. In some cases I can
> find the source as the value of the function name.
>
> Given that these packages are released as GPL, oughtn't the unoptimized
> source be freely available, hopefully with comments? Am I missing
something?
> Is there a central place other than mirrors where such source is retained?
> Sourceforge?

I'm not a lawyer but as I understand the various versions of the GPL the
requirement is that that source code be made available to whom ever asks
(though we all seem to take that to mean that the software is provided  in
source form). The provider may request coverage of costs for distributing
the source code; I vaguely remember that such a charge is to be "at
cost"
rather than prohibitive but it has been many years since I last bothered to
read the text of the GPL. Also I don't ever remember reading a clause in the
GPL as to the source being commented at all.

CRAN may impose additional criteria over and above those of the GPLs. Again
I don't recall reading any such criteria being imposed.

I'm assuming from what you wrote about value that the zipped files contains
an executable version of the package that you can use. Anything else is
surely a condition of the GPL but the provision of source with the package
not a requirement.

Regards, Trevor.

<>< Re: deemed!

	[[alternative HTML version deleted]]

On Jun 27, 2011, at 7:24 AM, Galkowski, Jan wrote:
> I wondered if there were standard practices in CRAN for delivery of R
source implementing functions in R packages. I has encountered a couple of
packages where the gzipped version of source contains very little, primarily the
Help files describing the functions in the package. In some cases I can find the
source as the value of the function name.
> 
> Given that these packages are released as GPL, oughtn't the unoptimized
source be freely available, hopefully with comments? Am I missing something? Is
there a central place other than mirrors where such source is retained?
Sourceforge?
> 
>  - Jan

Can you name the packages in question? 

Within CRAN packages (specifically the .tar.gz package source archives), there
will be several sub-directories which contain all sources files. These include:

'R' which contains all R source code

'src' which will contain any C, C++ or FORTRAN code and header files

'inst' which will contain other source code, such as Perl, etc.


If you are looking at the .zip or .tgz files, these are pre-compiled binary, not
source archives, to enable Windows and OSX users to install the packages without
needing to have development tools (compilers, etc.) present on their systems.

Regards,

Marc Schwartz

On Mon, Jun 27, 2011 at 1:24 PM, Galkowski, Jan <jgalkows at akamai.com>
wrote:
> I wondered if there were standard practices in CRAN for delivery of R
source implementing functions in R packages. I has encountered a couple of
packages where the gzipped version of source contains very little, primarily the
Help files describing the functions in the package. In some cases I can find the
source as the value of the function name.
>
> Given that these packages are released as GPL, oughtn't the unoptimized
source be freely available, hopefully with comments? Am I missing something? Is
there a central place other than mirrors where such source is retained?
Sourceforge?
>
The 'package source' link on CRAN should point you to a tar.gz file
that contains the source code. For example, for splancs off the heanet
mirror it is:

http://ftp.heanet.ie/mirrors/cran.r-project.org/src/contrib/splancs_2.01-27.tar.gz

 .tar.gz files from those links should have full R, C and Fortran source code.

I think we need counter-examples...

Barry

No, you are correct. It meets the letter of GPL. Took me a while to find
FunctionINIT.R though. As I wrote in the original, source is available, if only
by keying in function names and seeing their value. Was hoping for greater
clarity.

I of course have the paper and the help. I was trying to understand what this
"eta" partameter was and how to interpret it.

As mentioned, not the first package this has been an issue for.

Thanks,

   - Jan

- Jan, from Sierpinski, 
     a Blackberry, 6072391834,
     Google Talk to: 
       bayesianlogic.1 at gmail.com

----- Original Message -----
From: Gavin Simpson <gavin.simpson at ucl.ac.uk>
To: Galkowski, Jan
Cc: Barry Rowlingson <b.rowlingson at lancaster.ac.uk>; r-help at
r-project.org <r-help at r-project.org>
Sent: Mon Jun 27 11:36:57 2011
Subject: Re: [R] Standards for delivery of GPL software in CRAN packages

On Mon, 2011-06-27 at 11:14 -0400, Galkowski, Jan wrote:
> Fine.  Attached. It's "waved".
> 
> All it has is *.Rd files. Apparently the functions are collected in
> "functionINIT.R". But "00Index" and
"DESCRIPTION" are not helpful.
> 
>  - j
The Rd files are the help or manual "pages" for the functions defined
in
the package. In this particular case, the package author has decided to
put all the R code for their package into a single R source file -
functionINIT.R. The other two files are R-specific files, the latter of
which is used to describe the package; which, incidentally, points you
to a peer-reviewed paper that the package is support for.

I don't recall the GPL mentioning anything requiring that the source
code be "helpful". The authors have most certainly fulfilled their
requirements under GPL, as has CRAN in distributing the package sources.

Or am I being obtuse and completely missing your point?

G
> -----Original Message-----
> From: b.rowlingson at googlemail.com [mailto:b.rowlingson at
googlemail.com] On Behalf Of Barry Rowlingson
> Sent: Monday, June 27, 2011 10:18 AM
> To: Galkowski, Jan
> Cc: r-help at r-project.org
> Subject: Re: [R] Standards for delivery of GPL software in CRAN packages
> 
> On Mon, Jun 27, 2011 at 1:24 PM, Galkowski, Jan <jgalkows at
akamai.com> wrote:
> > I wondered if there were standard practices in CRAN for delivery of R
source implementing functions in R packages. I has encountered a couple of
packages where the gzipped version of source contains very little, primarily the
Help files describing the functions in the package. In some cases I can find the
source as the value of the function name.
> >
> > Given that these packages are released as GPL, oughtn't the
unoptimized source be freely available, hopefully with comments? Am I missing
something? Is there a central place other than mirrors where such source is
retained? Sourceforge?
> >
> 
> The 'package source' link on CRAN should point you to a tar.gz file
> that contains the source code. For example, for splancs off the heanet
> mirror it is:
> 
>
http://ftp.heanet.ie/mirrors/cran.r-project.org/src/contrib/splancs_2.01-27.tar.gz
> 
>  .tar.gz files from those links should have full R, C and Fortran source
code.
> 
> I think we need counter-examples...
> 
> Barry
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
-- 
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
 Dr. Gavin Simpson             [t] +44 (0)20 7679 0522
 ECRC, UCL Geography,          [f] +44 (0)20 7679 0565
 Pearson Building,             [e] gavin.simpsonATNOSPAMucl.ac.uk
 Gower Street, London          [w] http://www.ucl.ac.uk/~ucfagls/
 UK. WC1E 6BT.                 [w] http://www.freshwaters.org.uk
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%