What Larry Wall is doing (or has been doing, for the last seven years) with Perl 6 is revolutionary:

There needs to be a universal root language, and ways of warping that universal root language into whatever dialect you like.

It’s incredibly valuable to examine all the ways in which a (programming or natural) language can vary. I’m sure they’re learning a lot that will be useful to language designers everywhere. Perl 6 will undoubtedly be a great tool for quickly prototyping new languages, too.

It’s also wrong for a production language. The last thing I want in a programming language is diversely variable and configurable syntax and semantics. Sitting down to maintain or debug someone else’s Perl 6 sounds like a nightmare. Not only would you have to figure out what they are trying to say, you’d have to figure out what dialect of Perl 6 they were using, and exactly how that dialect works.

For this reason I’d tend to avoid Perl 6 for new projects, nor would I want to join a Perl 6 project mid-stream. If I had to work in Perl 6, the first thing I’d do would be to convince the entire team to stick to the standard dialect (or maybe some other, popular dialect) and never re-configure it.

I doubt we’ll see much traction for Perl 6 outside of research settings. The benefit of picking PHP and Java over Python or Ruby is that it’s easier to find PHP and Java programmers, than it is to find programmers for even the relatively popular Python or Ruby. Trying to find a programmer who knows not just Perl but Perl 6 and your dialect of Perl 6 is going to be even harder.