The jamsession package is based upon the notion of central storage of R sessions, R objects, and R scripts, to facilitate re-use and to minimize the time spent task-switching across analysis projects.
The methods subscribe to my personal priority for writing R functions, at least make my own life easier! These functions make my analysis life more efficient, as I know I can retrieve an R session from a project from several years ago. If I want to re-use an R object from an R session, I have a mechanism to do so that scales reasonably well. If I realize that I munged some data from a week ago, I can go back and reload my R session from a week ago.
The R session mechanism also changes the R prompt to include the project name, which I have found immensely helpful when multiple active R sessions are open, and I need to remember which is which.
The basic units in jamsession:
An .RData file based upon a project name, saved
alongside the Rhistory, in a central location. One can discover
a past R session with grep_jamessions("array950") and
load with load-jamsession(grep_jamsessions("array950")).
An .RData file which contains only one R object, intended
to be saved and possibly re-used in a separate R session. Objects
can be discovered using grep_objects("hg19refgene") and
loaded with load_object("hg19refGene").
An .R file which contains custom R functions related
to a specific project, but have not yet warranted being rolled into
their own R package yet. Such functions can be loaded or refreshed
with refresh_functions(project). The reason to store
functions separately and load via this mechanism is that the functions
can often encapsulate logic too complex to run in a live R console,
and to keep function code out of the R .GlobalEnv to save
space. The mechanism here will load functions into their own R
environment, whose parent environment is .GlobalEnv to allow functions
to be in the usual R function search space.