I wrote the Bioconductor package variancePartition (http://bioconductor.org/packages/variancePartition/) which is currently available for Bioconductor >=v3.2 and corresponds to R >=v3.2. I submitted a manuscript based on this software to a peer-reviewed journal, but a reviewer only has R v3.1 and was not able to install the software. Of course, the reviewer could have just installed a new version of R, but they incorrectly believed that v3.1 was the newest version available on OS X.
Is there any way to make my package available on older versions of Bioconductor? Obviously, this is not ideal, but it is currently a barrier to publishing
Yes, it is theoretically possible. If your package doesn't rely on anything else from the new ecosystem (base R features, or other bioc packages) then you could simply install the package using manual means (ie. how you might install any other package from source (not using biocLite))
But it can also open a whole can of worms which you shouldn't have to deal with. If I were you I'd just politely tell the reviewer that s/he is running a version of R that's > 1 year old, and trying to support that is more trouble than it's likely worth (for all parties involved)
Theoretically possible (in some cases) but completely not supported and very likely to cause problems. I would NOT tell your reviewer to install a package in this unsupported manner. They should upgrade R.
Theoretically possible (in some cases) but completely not supported and very likely to cause problems. I would NOT tell your reviewer to install a package in this unsupported manner. They should upgrade R.
Yes -- exactly what I said ;-)