True enough, Perl's unicode support is pretty good, but if your input is "encodable" in latin-1, it will still save you a lot of headaches to use latin instead of unicode if you work with remote applications like databases - DBD::mysql's utf8 support is still extremely unreliable for instance.

edit: Oh and also note that there is no standard for encoding characters outside the accepted charset: most popular browsers use the &#NUMBER; encoding, but not all of them do. See the replies to Unicode characters in <code> blocks for example