There are several systems used to categorize similar names so as to enable finding names that may have a similar sound but possibly a very different spelling. Two popular systems are the American Soundex system, which is used by the U.S. government on census data (and which is a variant of the Russell Soundex system), and the Daitch-Mokotoff (DM) Soundex system, which is especially useful for Eastern European surnames.

The DM soundex system is available here to help in finding surnames. No results will be returned unless the pattern is in ALL CAPITAL LETTERS.

Note that the implementation of the real-time DM Soundex system can be somewhat slow to execute, as it computes each time the soundex codes of every item in every database that it searches. (This is necessitated by the fact that the database contents are changing on an almost daily basis.)

The real-time DM Soundex was implemented in Java by a group of UC Irvine students. Here is the documentation and the collection of class codes.