I always forget that when you change the character encoding of a webpage (say from ISO-8859-1 to UTF-8 or UTF-16) that you also have to change the character encoding of the actual file.

We were having an issue today while working on internationalization where we properly configured our server to spit out UTF16 pages but all the external JavaScript and CSS were rendering out Chinese characters. We racked our brains for awhile until we realized the external files were saved with a text-file encoding of ISO-8859-1. Manually converting a thousand plus files to a different character set would have taken ages luckily, one of our other web devs, Bryan, found out that you could automate this via unix command line (piconv). Rejoice!

As an added note, be aware that switching from a single byte character set like ISO-8859-1 to a double-byte like UTF-16 will double your file size. Thankfully gzipping your files makes this final file size increase much more palatable.