We’re still catching up to Perl

Perl did a lot of things right a long time ago. For me a few of the big ones arent technical-- CPAN as you mentioned, was the most useful repository of any programming community. Another one for me is that Perl documentation has conventions and one of those conventions is a tiny "synopsis" at the top of each page that shows actual code using the thing that page is describing. I dont think people understand how much this has helped-- and its been that way for 20 years. Of course, in Java to show how any simple library works you'd have to show the contents of at least 8 separate files but hey- who's fault is that? ;)

Now that I think about it, the documentation in general (in Perl) was light years ahead of its time. perldoc, the included documentation is pretty comprehensive. The camel book came out when a lot of hacker news readers were a twinkle their father's eye. Perl was a big part of Oreilly's good ole days and it set Perl apart as a community that had really good accessible books available that werent CS textbooks but also werent "Learn how to programmer in 24 minutes".

I dont use Perl much these days. Most of my peers are afraid of Perl or were told to hate Perl. They dont know much about it other than when to laugh at a Perl 6 joke. A few point out that python has come of age and greatly out performs Perl. Still, there will always be plenty of perl in ~/bin.

I think in large part it's because Perl's formative years predate the blog revolution. Ruby, PHP and Java existed during Perl's ascendency, but they were in a gestational phase. When Ruby broke into the mainstream US programming culture was when blogging was at its height, and the open source advocacy culture that coalesced around Ruby on Rails became the new template. It was such a powerful influence that it even rubbed off on existing stacks like Java and .NET to some degree—these corporate-driven platforms actually needed a good dose of the simplicity and pragmatism that Rails and PHP were peddling. But for Perl they already had all the pragmatism, along with much better tools, and an established culture, so it probably seemed they had little to learn from the open source communities of the aughts.

If there's one thing that really hurt Perl in this new environment it was the unapproachability of Perl. It is clearly of an earlier era when it was taken for granted that programming was difficult and powerful tools with steep learning curves were favored. When I turned 18 in 1996 Perl really piqued my curiosity and I aspired to learn it, but what actually stuck was PHP simply because I was shipping so much faster. The whole second system syndrome with Perl 6 obviously didn't help either, but I think that's secondary to the poor on-boarding and general newbie-friendliness of the whole ecosystem.

Could you elaborate a bit about how you find TAP better than xUnit? I did some Perl testing years ago and mostly xUnit-style testing since then and it didn't strike me that xUnit was much worse than TAP. Genuine question since I'm very interested in the topic of testing.

Perl is not "old" culture. Sure, CPAN and much of the testing infrastructure including TAP is from the 90s, but so is Java. But a lot of the other cool stuff, such as Moose, is from the 00s.