question for ajith-parser comparison

Hi
I was wondering if there is any parser where we dont have to supply "ampersand(&) amp semicolon" (i am writing it out because the browser seems to convert it when the message is posted) to parse "johnson&johnson", basicaaly my question is....from the fron end we can ask our users to supply this so that our parser understands this(by the way we are using xerces.jar xalan.jar and crimson.jar for our application)....but when we reply back to our clients using the information in the database....we have to substitute each "&" with the appropriate "&semicolon" to parse the info back....is there any other way of doing this or is there any other parser which can do this for us
Thanks