hello all,

since japanese text commonly contains multiple scripts, what's the canonical way to romanize it? should i be running it through uconv -x three times, or is there a script-detection step which i have missed up to now?

thanks,
aaron.