I want to parse HTML using java, which way do you think is best? to use existing parser (which one is best) or to deal with the LL1 (top-down parser)?