>> Hadley Wickham <had...@rice.edu>
>> on Tue, 28 Aug 2012 10:16:31 -0500 wrote:
>> Could you *please* talk to some actual computer scientists and at least
>> consider calling it roxygen version 3.0 or roxygen2 version 3.0 instead of
>> releasing a new package name every time you feel the need to refactor the
>> code? It is commonly understood that major version numbers may break
>> backwards compatibility... no need to needlessly break the package name
too.
> I don't think you understand the reality of the R package management
> system. Most people will run update.packages() and get new versions of
> all packages. If a package has API breaking changes then it will cause
> considerable frustration, especially given how difficult it is to
> install a previous version of a package.
May be it's time to revert to "roxygen" name?
Roxygen is not required by any other package; so no dependence
problems. Also if user interface is not changed, and only internals are
refactored, then there is no need for a new name, is it?
Another option would be to release a new "old" package like
"roxygen_old", "ggplot_old" etc. Then people can just use the old one
with minimal inconvenience. The problem with this, is that there might
be papers/books published with the old command interface, but people
understand that, so not a big deal anyways.
Vitalie
_______________________________________________
Roxygen-devel mailing list
Roxygen-devel@lists.r-forge.r-project.org
https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/roxygen-devel
