Mostly because I don't have a compelling problem to apply it on. That's not quite true. I want to port Gopher::Server to using Perl6's object system (mostly because Gopher::Server has a nice design that should be ammenable to learning Perl6 objects). But Pugs doesn't seem to be entirely there with the objects yet.

Also, the latest releases need a more up-to-date version of GHC then I can get on my system ATM. GHC on Gentoo relies on Haddock, which appears not to be compilable with a hardened GCC (see http://bugs.gentoo.org/show_bug.cgi?id=92282).

In any case, Perl6 has the same problem that I currently have with other languages that are significantly different from most other languages I already know--it's too easy to say "I would like to solve this problem in a new language, but the old language will do fine and I'm having problems getting it to work in the new one".

The key is in how much langauges differ. After a day or two, I felt I knew C# well enough to tell a perspective employer that I know the language. I already knew Java, and C# is boaringly similar. I don't feel the same about O'Caml, which I've been playing with off-and-on for about a year.

Perl6 isn't as different from Perl5 as O'Caml, but taking advantage of the language is going to take a while.