You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A private rOpenSci slack discussion originally prompted by frictionlessdata/frictionless-r#120 and instigated by @peterdesmet let to @jennybc pointing us to tidyverse/design#42. The latter issue in particular raises some important points regarding general design for control of package-level verbosity.
Why is this important?
Because of current package-ecosystem-level options like tidyverse.quiet and component packages such as readr which now rely on package-level options to control verbosity. Any packages which ultimately, and not necessarily even directly, call potentially verbose tidyverse functions will generate output determined by such options.
Current practice
I'm pretty confident that running pkgstats over all rOpenSci packages - definitely including mine 😉 - would reveal a predominance of function-level control of verbosity through single parameters. Moreover, I'd be equally confident that the majority of such parameters would be logical, and so would not allow any finer-level control over verbosity.
Suggestions
Given the clear direction of discussion in tidyverse/design#42, we can likely assume progression away from function-level control of verbosity towards package-level control, along with more fine-grained options other than simple logical variables. I would argue that this alone suggests a clear need for some kind of "best practices" on control of verbosity.
The text was updated successfully, but these errors were encountered:
A private rOpenSci slack discussion originally prompted by frictionlessdata/frictionless-r#120 and instigated by @peterdesmet let to @jennybc pointing us to tidyverse/design#42. The latter issue in particular raises some important points regarding general design for control of package-level verbosity.
Why is this important?
Because of current package-ecosystem-level options like
tidyverse.quiet
and component packages such asreadr
which now rely on package-leveloptions
to control verbosity. Any packages which ultimately, and not necessarily even directly, call potentially verbose tidyverse functions will generate output determined by such options.Current practice
I'm pretty confident that running
pkgstats
over all rOpenSci packages - definitely including mine 😉 - would reveal a predominance of function-level control of verbosity through single parameters. Moreover, I'd be equally confident that the majority of such parameters would be logical, and so would not allow any finer-level control over verbosity.Suggestions
Given the clear direction of discussion in tidyverse/design#42, we can likely assume progression away from function-level control of verbosity towards package-level control, along with more fine-grained options other than simple logical variables. I would argue that this alone suggests a clear need for some kind of "best practices" on control of verbosity.
The text was updated successfully, but these errors were encountered: