tidyr and stringr are part of the tidyverse

[z3tt]

In your disclaimer you state that you

"use and recommend Hadley's (non-Tidyverse) packages ggplot2 and stringr" (emphasis added by me)

which are actually tidyverse packages.

Maybe something you want to fix (and changes the whole story as you are literally using the tidyverse -- and please do not argue that ggplot2 was around before the tidyverse and thus should not be part of it).

[matloff]

In your view, what is wrong with my arguing that way? It's true, and there is nothing "Tidy" about ggplot2. Indeed, HW has said he would design ggplot2 to be Tidy-ish if he were to do it today.

[z3tt]

It's wrong that these are non-tidyverse packages.

[aphalo]

@z3tt 'ggplot2' was designed several years before the tidyverse name or even the use of "tidy data" to refer to data organized in long form was coined by HW. 'ggplot2' is from the time HW was in academia and well before RStudio, both program and company, existed. The design principle is different. This is how I understand the point @matloff is making, he is considering the design principles, not how packages are bundled for distribution.

If you consider as tidyverse packages all those packages listed by tidyverse::tidyverse_packages() then yes, one could say that 'ggplot2' has been bundled into the tidyverse. However its design has nothing in common with other, newer, packages in the bundle. So maybe to avoid confusion, one could use "non-tidy" instead of "non-Tidyverse" to refer to such packages...

[z3tt]

I aware of all that, and your last sentence sums up nicely what this is about.