There are 100000 perl modules, all backed up by an extensive test suite. About jobs: knowing a popular language guarantees you a job, but it doesn't guarantee you a good job. Those shooting for a higher level can pick what language they want. Those looking for higher level programmers will find a much broader distribution of language preferences. Hence Twitter runs on Scala, Blekko runs on perl.
> Hence Twitter runs on Scala, Blekko runs on perl.
Those are hardly equivalent (aside from the obvious differences in prominence, scale and relative importance between Blekko and Twitter).
Scala shops attracts developers curious about new paradigms (FP / FP hybrid) whereas Perl shops tends to repel forward-looking developers who view it (accurately or not) as an ugly throwback to the year 2000.
Put plainly : Scala sells itself (to a certain crowd) whereas Perl needs to be be sold almost across the board. Scala is viewed as an up-and-comer, Perl as a has-been.
The parent was making reference to the common notion that using outlier languages provides access to high quality talent. In certain cases this holds true, in others not. You will attract, on average, higher quality talent by using Clojure rather than Cobol.
EDIT (response to parent edit) : well, no. I never said there are "no good Perl jobs", just that Perl is not in and of itself a draw in the way that Scala / Clojure / Haskell might be. You could probably convince someone that there are cool opportunities in Fortran development, but very few are actively seeking them out.
That thing about Perl is probably being said since the past 12 years and the needle doesn't seem to have the least count distance with regards to that.
People are learning Python? Really? That was probably true in 2006-2009. Python is no longer the new hotness. Scala and Clojure may be.
The fact of the matter is Perl has a massive niche(Text processing, Unicode, CPAN, Regular expressions, Fast prototyping, TIMTOWDI, Ability to add new syntax rapidly[Moose, Devel::Declare etc], Automation, Testing culture, Documentation, Unix, Parsing, Getting things done real fast... I can go on.. but I will stop here), If you look at what Larry Wall has been saying over the years about Perl 6 in various talks you will note that he designed Perl to fill a niche which continues to be totally unoccupied till today's date. The only competitors Perl has in its design space are tools like sed and awk with a additional few Unix utilities thrown around. In other words Perl doesn't have any credible competition. What's more? Any language that is designed to replace Perl will largely look like Perl(Read: Ruby, Perl 6). And yes, Php and Ruby do have their niches.
Python doesn't have a niche. All Python has is- "Tab indent your code." And sorry to break the bad news to you. Merely forcing indentation on a bad programmer doesn't magically transform him to a good programmer. Rather it make the situation more worse. It gives him a free pass to sit amidst good programmers.
> a niche which continues to be totally unoccupied till today's date
Leaving all about Python aside, what exactly is that niche? I'm not being ironical or anything, I really want an answer to this question, that's why I started this piece of the discussion. Seeing that things like Metasploit, that I would have expected to be written in Perl, end up being coded in Ruby, makes me unable to see any "special niche".
my theory is that Python ended of being used so much in scientific computing because of what it lacked :) I mean weird stuff that put off "brilliant non-professional programmers" (ie. scientists, engineers and other domain experts that thought themselves how to code). All other languages have tons on these things: Ruby has blocks that make people write functionalish code, Perl has ...tons of weirdness. Python just looked clean and easy to learn and lacked "weird shit" :)