Hi, (CC: Ricardo since they is the master about R stuff. :-))
On Tue, 09 Mar 2021 at 19:44, Mark H Weaver <m...@netris.org> wrote: > We have one notable exception in Guix: "r", which is "grandfathered in" > so-to-speak, but perhaps it would have been excusable anyway, given its > prominence and its large number of related packages "r-*". > > I fully support your proposal to talk about single-character names in > the manual, and also to check for them in the linter, but perhaps "r" > should be whitelisted. On the other hand, instead of simply the one letter “r”, we could rename it “r-environment”. Because the package currently named “r” is the interpreter named “r-minimal” plus extra packages, constituting the basic bricks of the “R environment” [1]. 1: <https://www.r-project.org/about.html> --8<---------------cut here---------------start------------->8--- (define-public r (package (inherit r-minimal) (name "r") (source #f) (build-system trivial-build-system) (arguments '(#:builder (begin (mkdir %output) #t))) (propagated-inputs `(("r-minimal" ,r-minimal) ("r-boot" ,r-boot) ("r-class" ,r-class) ("r-cluster" ,r-cluster) ("r-codetools" ,r-codetools) ("r-foreign" ,r-foreign) ("r-kernsmooth" ,r-kernsmooth) ("r-lattice" ,r-lattice) ("r-mass" ,r-mass) ("r-matrix" ,r-matrix) ("r-mgcv" ,r-mgcv) ("r-nlme" ,r-nlme) ("r-nnet" ,r-nnet) ("r-rpart" ,r-rpart) ("r-spatial" ,r-spatial) ("r-survival" ,r-survival))))) --8<---------------cut here---------------end--------------->8--- I also support the proposal. My point is: I am not even sure that “r” should be whitelisted. Cheers, simon
