For the example you mentioned, I would have used what's called "soundex". Soundex was developed a while back by god-knows-who (I can't remember and I hate it). It is a language somewhat simular to regular expressions, to determine the "pronouncement category" of a word.
SELECT soundex('jarred');
SELECT soundex('jarret');
SELECT soundex('jarrit');
All gives the same result, 'J630'.
However, soundex('hello') and soundex('yello') returns (respectivly) H400 and Y400. I'm not going into further details (search the web if you need them). Basically, you should either match the entire string, or cut out the leading character. This can produce some pretty fuzzy searches, but is NOT good for foreign languages (I have proven it not to be very compatible with norwegian ;-)
hint: SELECT SOUNDEX(field) FROM table
Read the mysql documentation, I think it has something about soundex.