DidYouMean's normalizing function could be split into different levels or could
take a "flags" parameter.

Levels might be:
*Capitals
*Accents
*Ligatures
*Extra characters such as spaces, hyphens, and apostrophes
*Comatibility characters such as curly vs straight quotes and apostrophes
*Similar looking but semantically different characters such as 0 vs O

Version: unspecified
Severity: enhancement