There are R packages for scientific papers developed around 2014. These packages are still working with its base code with C in the newest current R version and on CRAN. So I wonder for much longer-term projects, R is a better choice than Python?