Perl has a bright future, because it's already propagated everywhere. There's far more Perl in business environments than there ever was COBOL (though perhaps not proportionally :), so someone who can reverse engineer old transaction processing websites has a good, solid future.

I'll echo the comments about programming versus language, though; they're absolutely right. I think it's far better to have half a dozen Perl texts on my desk -- and to use them regularly while creating production code -- than it is to spend my waking hours mastering arcane syntax of things that <whisper> weren't implemented well</whisper>.

Use your time to learn new methodologies. I've heard good things about Higher Order Perl, but the Net is full of articles about the concepts. It will do your mind far more good to bend itself around conceptual leaps in Prolog or even SNOBOL in order to understand why they were built the way they were than it will to install a new Chatterbox client in order to socialize.

Programming languages WILL change. Our CPUs are so powerful now that we are the bottleneck. I see directions like IBM's Cell (multiple simpler CPUs on a chip) as becoming more popular. Is a linear paradigm a la Knuth still viable as the number of threads goes up? How much of our future will be parallel pattern recognizers and how much will be step-by-step iteration?

It may be that Perl will still be what it is five years from now, but the question will be, "How relevant is it?" At some level, and to some people, it will still be very relevant. FWIW, as long as we still use keyboards and alphabet-based language systems with low-level (iterative ALU-based) hardware, Perl will be able to contribute.

I believe, however, that we will move towards a future in which CPUs and memory are embedded in everything, large quantities of them, and the communication and prioritizing of their results will outweigh the importance of any given computation on any given CPU. The whole paradigm of creating a complete "program" and handing it to a CPU will become a much less important activity, I think, unless you are a systems implementor creating the underlying infrastructure. The activity of programming will be much more one of juggling capabilities of already existant nodes to accomplish tasks in a far more transient way. Imagine, for instance, that you have a sea of little CPUs out there, each of which has its own CPAN and its own action capabilities. Some will be drivers for picture walls, others will be listening nodes for spoken input, still others will run the environmental systems for your house. The API for those libraries will become much more important than the base language, IMHO, and it'll become a lot harder to tell the difference between code written for that purpose in shell script or in Perl. I think something more like Expect will become important, though Perl will probably mutate in that direction.

All that smoke blowing aside, becoming adept in Perl is a good thing, especially if you use it regularly to do real work. I'd take care, though, especially at your age, not to think that you have a plan that you can get comfortable with. It sounds to me like you're in a former Soviet satellite, yes? IMHO, your life opportunities may change a lot faster than Perl will, and keeping your antennae fully extended in THAT arena will do you much more good than mastering multidimensional references in Perl.