|Perl: the Markov chain saw|
We should have a star "blogger" (I hate that word, but it's the one to use I guess) that talks about Perl and the exciting things going on with development and in the community.
The problem is that we'd have to have a way to get people to read the blog. That means referring them from places where other languages have the excitement currently.
So many other bloggers use their blogs (some even running Perl-based software!) to show what they think are the better-than-sliced-bread features of other languages, all the while playing Perl5 down as a has-been. That's a lot to overcome in that arena. Yet, it's where the buzz for other languages is. From Ruby (37signals) to Python (ESR, among others), to Arc (Paul Graham), people hear great things about languages from the people they admire. They then take that as the truth.
chromatic and a few others are evangelizing Perl6 already on places like http://lambda-the-ultimate.org. There seems to be some interest in Perl6 even from such academic corners. However, among the people who discuss differences among languages as an academic exercise or as a hobby, the consensus seems to be that Perl6's strengths have a lot of negatives from Perl5 to overcome. The course of action I see these people taking towards Perl6 is to consider it a Lisp with a strong community of "those imperative programming people" writing useful libraries. They'll turn their noses up at our idioms, and the other people in their functional circles will snicker at them for using a language that can be functional but doesn't enforce that. Getting a few more world-class functional programmers interested in Perl6 isn't a bad thing, but this isn't the core of converts as I see things.
The question when you want to evangelize Perl5 is, "Who does Perl5 help do things the way they are already doing them, but better?". C programmers are one, obviously. People using Rexx, awk, sed, bash, and a thousand other pre-Perl5 languages have largely already moved to Perl. Sure, people still use C when Perl's not appropriate and C is. Some things are just simpler and much lighter-weight in a shell. I for one haven't touched awk, sed, or Rexx since learning Perl, though.
One of the biggest complaints against Perl seems to be its quirky OO system. If Moose is what lets people writing in Python or Ruby write clear OO code faster and easier with better maintainability and decent performance, then push Moose. Don't push Perl5 to people who have been indoctrinated that their language is better, specifically, than Perl5. Push what is new and powerful that addresses the negatives they've heard about Perl5. Hell, call the language Moose when you evangelize about it. Don't tell them it's Perl5 with a bag on the side. Tell people it's a Perl-descended language that sits on top of Perl5 as an implementation detail, and that future Perls will be using a similar object system. Either description could be considered equally valid from the proper point of view, but I know which one would sound better if I'd been indoctrinated against Perl5.