I wrote this once and it didn't post (that often happens here, in various scenarios) or else it got deleted. Assuming the former. Here's a super short version of the same thing. The first go was much better. Dammit.
* College CS departments are owned by Microsoft and Sun. They use C# and Java out of consideration for strings-attached grant money.
* Highschool and gradeschool kids learn PHP, jquery/JS, Squeak, Processing, Flash, or Python/PyGame. Perl has a foot in the web world (though not the appeal to the ultra-low-brow user base) but virtually no foot in the playing with graphics department. SDL is okay but it isn't easy enough or flashy enough to compete.
* Perl teams are small. Fewer people are hired to do a project compared to Java -- radically fewer. And companies hire almost exclusively senior level Perl programmers. It's hard to get a toe hold in the industry. You virtually have to publish a lot of great stuff on CPAN to get a job. Strong typing in Java I think makes it easier to integrate weaker programmers into a team and keep them from doing damage. They aren't kept out of your living room with a shutgun but with locked doors. Compartmentalizing the code, it's harder for someone new or intermediate to do damage; they just do or don't succeed. That's a big improvement.
* There is no perception that there's a lot of money to be made writing Perl. Java jobs with far lower expectations have paid me personally far better than Perl jobs with high expectations. Really good Perl programmers don't get actively recruited away. Countries with developing economies aren't tooling up on Perl and Perl companies aren't sponsoring H1B visas. Nor are Perl companies making long term investments in employees and letting them spend years only marginally productive with the idea that they'll be there for 10 or 20 years. Perl jobs just tend not to be that "milk". We all have just one battle story after another complete with scars. The lack of promise of money plus comfortable employment does not draw in adults in over Perl's job market.
* Relatedly, no one is out there just making cool stuff in Perl and saying, "Hey, look what Perl can do". Yahoo! Pipes is Perl and it's way awesome but they aren't playing up the Perl bit and there are precious few examples of this. People's perceptions would be that Perl is only used for big, cranky, serious old web apps. And they'd mostly be right.
* Perl was and perhaps still is the language of choice for system administration on Unix but other things are competitive in this field now. This is probably the largest avenue by which people discover and learn Perl -- Unix admins. Shops that do a lot of Unix administration probably still take non-programmers and tell them to learn Perl.
* Microsoft, Sun, IBM (SAP really wins here) are buying their customers lunch and being buddy-buddy with them. They're listening to them, agreeing with them, sympathizing with them, and drinking with them. In this regard, Perl doesn't even exist. Microsoft and Sun get a lot of money from companies but they don't just walk over with their hand out -- they woo them. This is a bonus point that's unrelated to the main point.
Want to kill something? Cut the supply lines. There's precious few new projects using Perl (neither large business nor garage style maybe-the-next-Twitter type). The avenues by which people have discovered Perl in the past have almost entirely been closed off. Nothing dies quickly. COBOL is still around. COBOL is even in demand -- the dearth of new people learning for the amount of legacy code out there itself creates demand. Except for system administration (I don't know, what else can you think of? Where do people come from?), our lines have been cut.
As I wrote in the last post, Perl is still assimilating and the community adapts very quickly after assimilating something. In about a year, half of the web infrastructure made for Perl switched to Plack. That's almost over night. Coro has been making waves. We ran with ORMs like tweaked out squirrels with scissors. It's disgusting, really. Collectively, we're having a lovefest with git.
Yeah, there's CPAN. Trying to figure out what happened to sunsite.unc.edu (go find out! You'll like the answer) I happened into CTAN -- the Comprehensive Tex Archive Network. No one gives a flying fuck how much stuff is in there because they have no plans to use Tex. CPAN isn't going to sell people on Perl. It did sell them on the idea of creating repo archives, though.
I hope you've enjoyed my little rant. My goal isn't to kick Perl while it's down nor is it to pointlessly piss people off. The subject is subjective and debatable and I don't mean to try to "win" the debate, only to add usefully to it. Past a certain point, saying "Perl isn't dead!" just is not constructive. We have a lot of work to do.
-scott
