I'll back up what tachyon said about Text::Soundex, with one caveat: it depends on your data. I looked into this for something at work, but as we are a medical company, some of the terminology is ... how do I put this ... not phonetically straightforward. There are a lot of Greek and Latin chunks in our terms, sometimes Greek and Latin and English components are intermingled in the same word. And English is still more or less a Germanic language. Sorta.

If you are dealing with last names, as your example implies, those can vary even more widely. I would suggest that you familiarize yourself with the data, if possible, before you make a decision, but 1) we all know to do that whatever the project, and 2) it isn't always possible.

--
tbone1
Ain't enough 'O's in 'stoopid' to describe that guy.
- Dave "the King" Wilson