What is the philosophy behind the workspaces in R?

Question

When I start R session from some directory, R automatically loads the corresponding workspace (if it exists). After I finish to work in this workspace I can decide if I want to modify (save) the current workspace. This logic is simple and clear.

What I do not understand, is what happens if I start R from some directory and then change the working directory by setwd(). As far as I understood the workspace corresponding to the new working directory will not be "loaded". I still see the variables and history from the previous working directory. Why?

Second, when I quit() R, I replace the work-space image corresponding to the "new" working directory by the workspace corresponding to the "old" directory. Do I interpret the behavior correctly? What is the logic behind this behavior? Can I switch to another work-space from R session?

Answer 1

Workspaces are stored in .RData files and are automatically loaded from current working directory when you start R. But working directory itself (and setwd() function that sets it) has nothing to do with workspace. You can load any workspace by explicitly specifying any .RData file:

load("c:/project/myfile.RData")

or

setwd("c:/project/")
load()