Entering edit mode
Now that the version of R no longer determines the version of BioConductor, I need to change my definition and approach to staying "current" with my institute-wide installation of R and attendant libraries, especially BioConductor. I used to promise to upgrade both R and libraries 1 month after R version release, which ensured I stayed reasonably current with both. I also kept the old R available, available from command-line as, say, `R-2.15.0`. Now, I am faced with two releases of BioConductor library for each release of R. I am tempted to name my R installation after the combination of R and BioC version numbers, say, R-3.1.3-BioC-3.0.0. and promise to upgrade within 1 month post BioC release.
Dragons lurking?
What are other's doing?
Thanks Hans.
Can you tell me what the three stars are in `R_LIBS_USER=/***/***/3.1.0-bioc-2.14-release/`?
Are you suggesting that I keep two separate library installation locations, for example, one for BioC-2.14 and one for BioC-3.0, and the let the user pick between them? That might work.
Hi Malcolm
I just used the three stars to hide the actual path on our system - I once learnt: for security reason, never to post the full path ;)
Yes, I strongly suggest to keep two (and over time thre four,five, etc) library installation locations. It makes it much easier to reproduce your stuff. And if space is an issue you can always move the older installations away from the high-performance (expensive) storage to cheap storage.
Hans-Rudolf