This is a very open question, and it's not very well defined (you say '50% of the letters', but "McGinley" and "O'Griffin" both have 8 letters, but share only 3). A discussion of what is a good match and what isn't is not Perl related, so I won't go into that. However, you might want to look at the 'Text::Soundex' module, and see whether that will be of any help to you.

Why this modules comes with the standard distribution is a mystery to me, but it makes it easy to check out.

Abigail