[Bioc-devel] Size limits of packages when including other executables

On Thu, Jun 6, 2013 at 1:48 AM, Thomas Dybdal Pedersen
Hi

I'm developing a wrapper for the peptide-identification tool MS GF+. The algorithm is developed in Java and the .jar file has a size around 20 mb.

For the ease of the user, I think it would make sense to pack the java code together with the wrapper (this has been cleared with the MS GF+ developer), and would have the added benefit of securing version compatibility. This, on the other hand, stretches the size limits imposed in the guidelines by several factors?

What is the opinion on this? Is the size limits set in stone or can they be lifted for certain cases besides annotation packages?

We generally do frown on this, you're right, but if you're sure all
the classes in the JAR are necessary, then we'll allow it.

What's more important than getting buy-in from the developer of the
Java code is finding out what license the Java code is written under,
making sure that is compatible with the license of the R package, and
acknowledging the Java license as well as the R license in your
package. See the licensing section of the "Writing R Extensions"
manual for more information. Note that some licenses require that all
source code be either in the package, or that a link be provided that
points to it; this would apply to your jar file (which presumably
contains only .class files and no Java source).

Dan
best

Thomas Pedersen, PhD student at DTU, Denmark
_______________________________________________
Bioc-devel at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
Most of the content of the jar-archive is model parameters, so the actual classes doesn't take up much space. Thanks for the pointer to the licensing issue - will make sure that everything is sorted out?

best

Thomas

Den 07/06/2013 kl. 22.50 skrev Dan Tenenbaum <dtenenba at fhcrc.org>:
On Thu, Jun 6, 2013 at 1:48 AM, Thomas Dybdal Pedersen
<thomasp85 at gmail.com> wrote:
Hi

I'm developing a wrapper for the peptide-identification tool MS GF+. The algorithm is developed in Java and the .jar file has a size around 20 mb.

For the ease of the user, I think it would make sense to pack the java code together with the wrapper (this has been cleared with the MS GF+ developer), and would have the added benefit of securing version compatibility. This, on the other hand, stretches the size limits imposed in the guidelines by several factors?

What is the opinion on this? Is the size limits set in stone or can they be lifted for certain cases besides annotation packages?

We generally do frown on this, you're right, but if you're sure all
the classes in the JAR are necessary, then we'll allow it.

What's more important than getting buy-in from the developer of the
Java code is finding out what license the Java code is written under,
making sure that is compatible with the license of the R package, and
acknowledging the Java license as well as the R license in your
package. See the licensing section of the "Writing R Extensions"
manual for more information. Note that some licenses require that all
source code be either in the package, or that a link be provided that
points to it; this would apply to your jar file (which presumably
contains only .class files and no Java source).

Dan

best

Thomas Pedersen, PhD student at DTU, Denmark
_______________________________________________
Bioc-devel at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel