On 08/21/2017 06:39 PM, carl fakhry wrote:
Hi Martin,
I wrote an R package which has functions that invoke system commands. Part of
the system commands which I am using is calls to the ViennaRNA package
(https://www.tbi.univie.ac.at/RNA/#) and calls to python which requires a
specific python library, Biopython. Both of these packages are very commonly
used in the bioinformatics world.
The R package can install properly without installing ViennaRNA or Biopython.
However, to use the package the user would need to install them to have the R
package work properly. What should I do in this case? Should I make an extra
vignette which details the installation process?
Hi Carl --
Actually, if the R package is only useful with the relevant software,
then I think it should fail to load, with an informative error message
(implment .onLoad <- ... in R/zzz.R, for instance).
While a vignette is 'nice', I think an INSTALL file in the top-level
directory is more appropriate (one will often be trying to convince a
sys admin to install this additional software, and they appreciate plain
text); your error message could point to that (e.g., on bioconductor,
at
http://bioconductor.org/packages/release/bioc/install/ChemmineOB/INSTALL).
You need to carefully justify the use of non-R software, in particular a
lot of Biopython functionality is readily available in Bioconductor, so
it wouldn't make sense to add a fragile interface and additional system
dependencies to duplicate available functionality.
It is also important that the software, including INSTALL instructions,
follow very standard protocols on Windows, Mac, and Linux platforms --
the Bioconductor build system in particular needs to be configured to
work with all packages using system dependencies, and some combinations
will be very difficult for us to implement or maintain.
You'll also need to verify, more so than a typical Bioconductor package,
that your software works across platforms. Things like paths, temporary
directories, and encodings are all likely to be particularly problematic.
And don't forget to verify that the licenses of these software packages
are consistent with your own license, and do not impose odd
restrictions, e.g., for academic use only, that would make the R package
unsuitable for Bioconductor.
Martin
Thanks,
Carl
[[alternative HTML version deleted]]
_______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
This email message may contain legally privileged and/or...{{dropped:2}}
_______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
