Those solutions assume text is tokenizable, and suck at regex. I think they are the wrong tool for anything having to do with domain search. Furthermore, they don't play nice with continuous cron indexing.