First, even if perl 6 was released tomorrow (and trust me, it won't be) that's absolutely irrelevant--it'll be at least six months before the code has been beaten down enough to work out the bugs and get it in what I'd consder a production stable state. (I may have higher standards than other people) And, since you'll probably want another six months or so of small projects to work out the changes, it'll probably be a year before you'd want to consider it as a candidate for production code. (though the shaking out of bugs and playing with perl 6 time may be coincident)

Second, perl 5 isn't going away. There are many places that have perl 5.6.x, 5.005, and even 5.004 in production and active development. 5.004 is not new, and neither is 5.005. The version of perl you choose now will be fine for several years to come, and there will be new versions of perl 5.

Third, there will be a coversion tool, as well as a perl 5 front end to parrot, perl 6's back end. Who knows, given development efforts, we may well have parrot be a darned close to complete perl 5 engine before a perl 6 engine. Odder things have happend. (Though, for a variety of reasons, you won't see a parrot-backed perl 5)