Hacker News Reader

Comment by chunkyks | Hacker News Reader

Is tidyverse still fundamentally license incompatible with everything? Last I checked, it included both gpl2-only, and gplv3+, modules. That prevented my company from releasing code that we wanted to.

(this is really a pervasive problem across the r ecosystem, but tidyverse really seals the deal because it's so popular)

View on HN

We are in the middle of a process to systematically re-license as much as possible as MIT. More on that soon.

I had noticed some threads about this on stringi/r. Glad to hear its more systemic!

Thanks, Hadley. I really just wanted to say you're amazing, on the off chance nobody has told you that today.

The good news is I don't think there is a feature in the tidyverse you couldn't do with base R, albeit in an "R way." I think the main sell of tidyverse is the sexier plots from ggplot2 and some quality of life functions. Maybe others can comment if there is more there. The annoyance of 85% of stack overflow R comments starting off with sourcing the tidyverse library was enough to push me away and develop a base R solution whenever I had an issue.

I think the appeal to the tidyverse "universe" is the readability of the code. Especially to novel users of R (and there are a lot of people in science that have to generate a plot once a month or so). ggplot2 works fine if you only use that package without the rest of tidyverse. I personally use the data.table package a lot and than generate plots using ggplot2. Stackoverflow is indeed very much focused on tidyverse packages which is a pity as indeed base R can solve so many problems posted too without the need of additional packages.

That's the thing of it, too. It's pretty looking code, but it's not R. It's its own foo you now have to spend time learning on top of R.

Are you bundling dependencies? In the R ecosystem you can just make users download libraries themselves.

R itself is GPL2/GPL3: https://cran.r-project.org/doc/FAQ/R-FAQ.html#Can-I-use-R-fo...

I think that a compiler or interpreter using GPL does not affect you in the same way as a library.

Right. I'm not sure there's ever been definitive rulings on this but the viewpoint in the R community has been that it is the licensing is at the package level in terms of what a user would care about. As an example, in the early days there was an intentional decision by the Bioconductor group to use more permissive licenses in order to enable corporate use and thus help with adoption.

A quick look and all the packages are MIT except for purrr which is gpl2.

There's a mix of GPL-2, GPL-3, and MIT.

You couldn't release your code under GPLv3?

Reply on news.ycombinator.com