Dependencies, archivals, and package development in the R ecosystem
Tad Dallas, Michael K. Borregaard · BioScience · 2025
Abstract The accelerating rate of statistical package development has allowed researchers to easily perform increasingly advanced statistical analyses and has enhanced engagement of researchers in both writing analytical code and developing packages. However, this increase in the number of packages forms a correspondingly complex web of package interdependencies. This web aligns the many packages into a coherent ecosystem and greatly facilitates the ongoing development of new functionality, but package dependencies are not necessarily coordinated and, as , such also represent potential points for code to break. In the present article, we explore temporal trends in R package development, highlighting the rise in packages, dependencies, and archivals (as a way to quantify breaking packages). Following R Core Team developer Kurt Hornik's question from over a decade ago, we wonder “Are there too many R packages?”