[Rd] [RFC] A case for freezing CRAN

Jeroen Ooms jeroen.ooms at stat.ucla.edu
Wed Mar 19 23:16:41 CET 2014


On Wed, Mar 19, 2014 at 2:59 PM, Joshua Ulrich <josh.m.ulrich at gmail.com> wrote:
>
> So implementation isn't a problem.  The problem is that you need a way
> to force people not to be able to use different package versions than
> what existed at the time of each R release.  I said this in my
> previous email, but you removed and did not address it: "However, you
> would need to find a way to actively _prevent_ people from installing
> newer versions of packages with the stable R releases."  Frankly, I
> would stop using CRAN if this policy were adopted.

I am not proposing to "force" anything to anyone, those are your
words. Please read the proposal more carefully before derailing the
discussion. Below *verbatim* a section from the paper:

To fully make the transition to a staged CRAN, the default behavior of
the package manager must be modified to download packages from the
stable branch of the current version of R, rather than the latest
development release. As such, all users on a given version of R will
be using the same version of each CRAN package, regardless on when it
was installed. The user could still be given an option to try and
install the development version from the unstable branch, for example
by adding an additional parameter to install.packages named
devel=TRUE. However when installing an unstable package, it must be
flagged, and the user must be warned that this version is not properly
tested and might not be working as expected. Furthermore, when loading
this package a warning could be shown with the version number so that
it is also obvious from the output that results were produced using a
non-standard version of the contributed package. Finally, users that
would always like to use the very latest versions of all packages,
e.g. developers, could install the r-devel release of R. This version
contains the latest commits by R Core and downloads packages from the
devel branch on CRAN, but should not be used or in production or
reproducible research settings.



More information about the R-devel mailing list