I'd like to be able to modify the text passed to the parser for embedded Java code, to handle a character escaping system for my custom language. Specifically, it would be great to be able to pass a series of tokens to a single parser, and possibly convert those tokens to different textual representations in the middle. That way, my parser could parse through the string, and pass the tokens to the Java chameleon, and convert things like \n to real newlines before passing them to the chameleon. I think this is a lot to ask for, but I think it will help other languages, not just my own. For example, Javascript could remove that CDATA hack in the lexer. Also, it could handle XML entities like &nbsp as the browser will handle them - this would be incredible because I don't think any HTML editor currently handles these correctly.