First of all, thanks for an awesome project! 🏅

The current implementation will not complain about unused method arguments if the method raises a NotImplementedError (it was implemented in this PR: #7739).

However, many recommend against raising this exception in your code:

• https://oleg0potapov.medium.com/ruby-notimplementederror-dont-use-it-dff1fd7228e5
• https://blog.airbrake.io/blog/ruby-exception-handling/notimplementederror
• http://chrisstump.online/2016/03/23/stop-abusing-notimplementederror/

These three posts all boil down to:

1. Its actual meaning is different than the name implies – according to RubyDoc it should be raised when a feature is not implemented on the current platform, think of low-level things like calling fork or fsync.
2. It's a descendant of ScriptError so a standard rescue block will not capture this, which may cause unexpected problems.

There is also ongoing discussion in this sibling repo to Rubocop:

• NotImplementedError should be prefered for abstract definitions ruby-style-guide#458

Downside

It's somewhat weird to have def read_db(_name), where the real method name is name.

Possible solution

• Add an option called IgnoredExceptions (array type), with the default value of ['NotImplementedError'].
  • For anyone who hasn't configured this cop, it's 100% backwards compatible.
    • People who have set IgnoreNotImplementedMethods to false would need to update their config to use IgnoredExceptions: [] intead.
  • Could then deprecate the current option called IgnoreNotImplementedMethods and remove it in the next major.
• Allow the yarddoc @abstract keyword to be used as another signal to disable this lint check on a method.
• Remove the exception case for NotImplementedError (if you think the reasons for doing so makes sense).
• Do nothing and consider this an edge case.

Docs for Lint/UnusedMethodArgument

Addendum

Also, I realize this is nuanced. So feel free to close this issue if you think it's irrelevant or that the current logic is "good enough".