Okay, I have some thoughts about this.

• First, the basic documentation that people read when they're learning Perl needs to be updated to at least suggest that people use certain modern things like Moose and Catalyst. There's More Than One Way To Do It, and that's great once you've got a feel for the language and can make decisions more easily, but when you're first learning something, it's nice to have somebody say "just go use this."

  I think this means a new edition of Learning Perl, and updates to the man pages that ship with perl. Possibly also, as some said above, a new Perl 5 version of the Camel Book.

• A lot of promotion of Perl as a serious language for major applications needs to be done outside of the Perl community. I love Perl Buzz and use.perl, but those are mediums that are read by people who already write in Perl.

  There is a perception that Perl is a write-only language, and that perception needs to be broadly proven untrue. Somebody needs to advertise some extremely-well-crafted Perl. Not some really clever Perl hack, but some really-well-designed huge application.

• New Perl users need to be treated kindly, not berated or insulted. I think the Monks are pretty good about this, but I mean on IRC, in other forums, on mailing lists, etc. It should become unacceptable in the Perl community to be rude to new users, even if you're the King Of All CPAN. Nobody should have the attitude of "I've been around a long time so it's okay that I'm rude."
• In general, any encouragement of existing CPAN modules to spiff up their APIs and documentation would be great. If we want to be treated like professional programmers writing in a serious language, we should present ourselves that way. I'm of the opinion that Java is so popular because (a) it has great API documentation (b) it has great tutorials and (c) it got into universities as the primary teaching language.

  I think a great TPF grant would be for somebody to spiff up the documentation of the most useful and popular Perl modules.

• Updating the web site would be good, too, as some people above have said. Catalyst has the right idea, there. :-) In addition to being nicer to look at, it should be explicitly extremely helpful to new users.
• An improved method for searching CPAN is probably needed. I think that CPANHQ could do this, particularly with its planned tagging feature.

Those are just a few things that I thought up off the top of my head.

Perl is not dead. Now is the time to promote and get serious about keeping it alive, not after it's already died. It's true that Perl is huge, but so was Rome. That might seem like a silly analogy to some, but I think it's very accurate. Rome was so massive and so obviously invincible that nobody worked to prevent its fall until it was too late. Perl is so obviously huge that perhaps some don't imagine that it could ever die, but just because something is huge doesn't mean it can't shrink and shrink and shrink until it hardly exists anymore.