This week I realised that a script I wrote about ten years ago has become an indispensable part of my company's infrastructure: it converts Office documents to PDF using OLE. It's about 30 lines long of very clear code.
perl5.6.2 still blows away all newer releases performance and memory usage wise. 15 years ago the author was still active.
stableperl (http://blog.schmorp.de/2015-06-06-a-stable-perl.html) is a fork of perl 5.22 with a couple of tiny reversions that re-enable access to some internal state required by certain CPAN modules.
The author talks in other blog posts (http://blog.schmorp.de/2016-12-29-griefing-the-perl-api.html, http://blog.schmorp.de/2015-11-12-tidbits-why-coro-crashes-o..., http://blog.schmorp.de/2015-12-15-tidbits-cgipm-a-data-point...) about the decline of Perl's cohesion.
It's sad to see.
--
EDIT: And now I'm being downvoted.
I have no problem with Perl, I watched the Perl 6 presentation when it was released with great interest.
Perl remains on my list of languages to learn (I know, I know...). It's up there. And that's why I said I was sad to see datapoints like the one I linked. I want Perl to keep going strong, and I hope it lasts another 20 years. I guess I added this comment because I wanted to find out if there was anything I could do.
But for smaller scripts that automate tedious tasks or massage data to connect two otherwise software systems or generate a report-ish overview of a database/table, it is very hard to beat. Ruby comes close, but then again, Perl has been relentlessly optimized and fine-tuned over the last ~15 years, so for the stereotypical use-cases it is actually very fast, even compared to what one could cook up in C without weeks of hand-optimization and tweaking.
Someone once said that Perl is to batch-processing of text what Emacs is to interactive editing. I think the comparison is somewhat accurate.