[R] tidying up

[Avi Gross]

This mailing list seems to steadily get messages that some see as not relevant to this forum. In particular, some see it as wrong to bring up some things here and keep reminding people of some ground rules.

So I want to know, briefly, if it is reasonable to ask a person with a question or problem to reproduce their problem another way. If using RSTUDIO or one of many IDE, can they run the code on a naked R interpreter by sourcing the file or copying it in or typing it anew, or perhaps using IDLE which comes by default with many installations of R? If using a library (which I like so I am not really in agreement about the unsuitability) like the tidyverse which is free and available to all even if loosely associated with the RSTUDIO folks and that can be run on any version of R that I am aware of, then some questions may still be fair if they are really about more general R issues as much of the rest of the code may be base R and may be the cause of whatever issue is being reported. And, some simple requests like pointing out a missing comma, ...

I have stated my thought before and it boils down to the reality that there are some things about earlier versions of R that were far from perfect or complete and a little healthy competition is not a bad thing and may help base R evolve. Not everything in the tidyverse is better and it keeps evolving and deprecating older features, but it cannot really be ignored any longer. If you apply for a job at some company as something of an R expert, you may well be asked by all kinds of people about their programs that use the tidyverse for help or to help them solve a problem. You don't have to like it, but if you cannot read it, you no longer are really qualified in many places.

Ask yourself if a language like R was created from scratch, what graphics might be built into the base distribution? Would you rather have lattice or ggplot or perhaps both as well as base R graphics? Would you make many of the built-in functions more consistent, so for instance, the data being worked on would be the first argument whenever possible? 

One reason there are so many packages is not so much due to the superiority of R but because people find it lacks quite a bit. Much of that should not be included, of course, if R is meant to be somewhat on the lean side, and yes, packages are a deliberate way to extend it when needed. But when people use it and think they are programming in R, ...