Unicode tools (at least should) have ways to determine visually similar letters. Maybe someone more knowledgeable about Unicode than me can pull out the term for it (maybe "homograph"?). For example: 'ö' and 'o' should 'match' using this method. It also allows you to do things like make sure that μ (mu) and µ (micro sign) match.
So in response to:
> I'm not sure if interchangeable is the right word. 'phởne' and 'phone' yield different result lists.
The results are different because it's matching somethings that are an exact character match, and others that are just visually similar (homographs?).