Just a quick thought on the new Unicode features, sorry for the lazy message, but has there been any thought put to Unicode normalization forms? ie, say one module uses a variable $Régie but the other one uses $Régie. You've also got the issue of Mac OS X filesystems which normalizes filenames to NKC. Sorry if this has already come up, I just came across a bug in a piece of Perl code and this stirred this thought. SamThread Next