Replies: 3 comments 2 replies
-
I think that breaking changes should make it clear to the user that things are broken-- even if they aren't paying attention. For that, I think that the simplest change would be to rename the column to "fraction". That way, if they have code looking for "percent", it will break and they'll find it quickly. I think it would be harder if the name stays "percent" and the value changes to be 100-times the current number. Then, code would break but in a silent way. The way to do that, in my opinion, would be to add an argument like Or a simple way (for the user and not for the maintainer) is to borrow the concept of editions from other tidyverse packages (I first think of |
Beta Was this translation helpful? Give feedback.
-
I agree with @billdenney that changing the value without changing the name is a bad idea in the current package. @sfirke You've talked before about separating the cleaning functions of janitor and the data summarization functions to create a second package for the latter. Perhaps rather than making a new edition, this change could be implemented when this new package is created. |
Beta Was this translation helpful? Give feedback.
-
A split does sound like the cleanest path. For that, perhaps something like the following would make sense:
From there, hopefully, it's only a few revdep checks for the CRAN packages (https://devtools.r-lib.org/reference/revdep.html). |
Beta Was this translation helpful? Give feedback.
-
People keep complaining that the column name "percent" is technically inaccurate, see #300 and other related issues. Should it be renamed? I had figured it wasn't worth a breaking change, but maybe it is? In practice I don't see how anyone is going to be confused, but if users are always renaming the column because they find it wrong, that's not great.
If it should be renamed, what is the best replacement name?
Beta Was this translation helpful? Give feedback.
All reactions