'R CMD check' is essential and instrumental to the success of #RStats & trust in results. Anyone knows the history of R's package testing? Was it there from the beginning of R? Did it originate from S? Was testing a core component from the early days at Bell Labs? Who to thank?
5
9
1
39
I asked Kurt and according to his memory: 1998: Kurt & Fritz Leisch started R CMD check as a shell script that ran the examples (as already shown by @RogerBivand). 2000: Fritz with support from Kurt & BDR rewrote the checks in Perl with extended functionality.
3
2
18
Replying to @AchimZeileis @henrikbengtsson @RogerBivand
That was my recollection too (which I sent to HB via DM). High-fidelity testing of the packages in the emerging repositories is a key driver in how #Rstats started to have code archives that mattered, and which grew tremendously because of this focus on ensuring code quality.

Apr 12, 2021 · 1:43 PM UTC
2
9