"... Perl has an best regular expression power. Also, it has a difficult to learn, but the best unicode support."

I was just about to make that point. Python's regular unicode handling sounds like it's getting better, but they're regexps still don't deal with it very well. I'm not sure about Ruby, but some years back Matz went with an encoding agnostic approach... if I've got it right, encoding is an attribute of a string, which makes me wonder what happens if you concatenate two strings with different encodings, one unicode and one something else...

(All of them, however, beat the current generation of Javascripts, which still use the completely obsolete UCS-2 internally...)