Have you looked at Perl's String::Approx module?<P>If it does what you want, and still isn't fast enough, the algorithm is already coded in C (for that module,) so you can probably just rip out ...