< a quantity that can be divided into another a whole number of time />

August 11, 2019

Ease of learning vs relearning, by John Cook. Nice points, as always. I have just some minor concerns with the last paragraphs where the author says that the tidyverse is great beacuse of its consistency. First, pending some minor annoyances with the naming convention of formal arguments in base R functions–and recommended packages–, which I always called R’s language idiosyncrasies, I do not find base R that much inconsistent. Second, I disagree with the idea that the tidyverse comes with that much conceptual integrity, for what I used to see. Most importantly, there are so many dedicated functions in, e.g., dplyr, that it goes against the principle of compositionality that we use to like in functional and scripting languages. Finally, what used to be available in a short number of packages, but especially base, is now scattered throughout several packages (forcats, glue, etc.), so that I have a hard time believing that newcomers could find their way as easily as they would with base R only. Anyway, that’s my 2¢, and it is nowhere a critic of Hadley Wickham’s account to the R ecosystem. #rstats