How to ensure that R / Rcpp code is reproducible (“redistributable”)?

I wrote some R-code for the dissertation, relying on some external packages (for example, plyrand reshape) and writing a couple of relatively simple C ++ built-in functions using inlineand RcppArmadillo.

I would like to make sure that it can be executed “as is” on computers other than mine (Win64), for reproducibility studies.

My question is: Suppose I have included the code to install the necessary packages, the packages would be RcppArmadillo(and Rcpp, and inline) sufficient to compile functions written in RcppArmadillo, or end user to change the way system to compile on my Windows machine? If not, is it possible / recommended to save the compiled functions from my end and include it in the R code I'm sending?

In addition, in the unlikely event that the code needs to be run after some time (say, after a couple of years), is it enough to include a complete installation of R with the appropriate packages in their current version to make the code "future" -proof "?

I hope the question is clear.

+5
source share
1 answer

, , "" " - ", . "", , , , ( ).

" R-" - .

, Rcpp, , Rcpp- . "Rcpp-package".

+9

All Articles