How to choose among .data, .x, and x when designing the signature of a function?

Has anyone articulated some principles to help me to decide if the signature of my functions should use .data, .x, or x?

If the input is a data frame, use data; if it's a vector, use x.

Add a leading . whenever the function also takes named .... If a high proportion of functions in a family take ..., you may want to use a leading . everywhere, just for consistency. (see https://github.com/tidyverse/principles/issues/18#issuecomment-428606329)

3 Likes

I had the same question which had some useful discussion here: Function argument naming conventions (`.x` vs `x`)

Thanks @torvaney. Do you think it's worthwhile adding a pointer from there to here? Or maybe collapse the two discussions?