In order to keep it flexible, we allow users to search for both upper and lower case (i.e. don't just convert the case) when searching for bad characters.
However, this should be made explicit by setting the default 'bad characters' as 'B J O U X Z b j o u x z' and adding a reference to it in the documentation.