Hacker News Reader

Comment by bhaavan | Hacker News Reader

This feels like a last ditch attempt to save a dying language. (https://i.imgflip.com/4676mf.jpg) I wonder what exactly does Perl bring to the table as a language, why would one consider choosing it over other languages.

View on HN

Regular expressions as a first class language feature.

You should see some of the atrocities I had to debug at aol because of regex abuse.

Ruby also has this.

Ruby really is Perl done right. I can't really see any argument for Perl over Ruby as a language.

use strict. In Perl you can prevent something like the following from compiling. You can't with Ruby (unless they've finally addressed that).

   $foo = 2;
   if (some_test()) {
      $fooo = 3;
   }

Shouldn't that be:

    my $foo = 2;

?

I was leaving the declaration implicit.

Speed. I moved mostly to Ruby years ago, but when doing lots of text processing, a trivial script in Perl runs a lot faster than a similar trivial script in Ruby.

People are interested in using it and people are interested in developing it, so it continues to be used and developed.

I got my start with Perl; I wrote a book about it, I spoke at every conference, I did training, I maintained the Emacs mode. I started working at Google and, forced to use static languages, found them to be just as productive as Perl. And, they eliminated a lot of problems with Perl -- libraries were resolved at compile-time, and I just ended up with a binary that could be run anywhere (no @INC madness); types were checked at compile-time, so silly errors that required extensive unit testing in Perl could just be automatically underlined by my editor and fixed before I even saved the file.

Something else that bothered me was how much mindshare the dynamic languages were competing for among themselves. I would prefer to write a Perl program over a Java program any day... but Perl did not seem to be fighting with the Javas and C++s of the world, instead it was always trying to take on Python and Ruby. And Python and Ruby were like that too -- every Ruby programmer was out to kill Perl because they didn't like Matt's Script Archive. (Nobody wrote 1990s Perl in the 2000s, so it was kind of a strawman.) You ended up with a bunch of like-minded people fighting for the small attention share of "we don't really care about runtime performance", and all the factions were just too small to take on the larger issue of entrenched static languages. Java and C++ never felt like they had to take features or ideas for Perl, Python, or Ruby. So it all seemed kind of pointless.

With all that in mind, I do think Perl failed to "win". It is a neat programming language, but probably too complicated for beginners and not productive enough for people that have gotten comfy with C++/Java/Go, which are pretty darn productive these days. Meanwhile, Python found its niche pretty much everywhere -- you can program microcontrollers with it, you can write a video sharing website with it, you can do data science with it -- and it's a great introduction to programming for beginners. And, Javascript kind of came out of nowhere to conclusively "win" the dynamic language war (because a dynamic language needs a runtime, and guess what runtime exists on pretty much every computer and phone around these days?)

TL;DR: I kind of agree that the programming language wars have been fought and lost by Perl. That doesn't stop people from wanting to make it better, or to continue to use what they know. You can make lots of great things with it, so people continue to use and improve it.

Perl lost the war because it didn't turn-up for the battle. In the 2000s it went into a kind of COVID19 self-isolation phase. First there was Parrot - a half-assed attempt to create a CLR for dynamic languages. When that failed the MoarVM adventure began. There were about 3 implemetations of Perl6 in development simultaneously (MoarVM, JVM & Rakudo) as Perl5 hemorrhaged mindshare. All-in-all it took 15 years to produce a 1.0 which was so dog-slow for Perl's common use-case it was unusable ... and still is.

Yeah, Perl 6 was certainly A Thing. I am not sure practicing Perl programmers cared much about it. I went to the talks and thought "neat" but it didn't really seem like "welp, this is taking forever, I'm switching to Python." Maybe if everyone working on Perl 6 decided to work on data science tooling instead, Perl would have "beaten" Python in that area and things would be different, but the world doesn't really work that way, so I don't think Perl 6 really mattered in terms of Perl fading away. A lot of people learned a lot of things about language design, and that will no doubt benefit the entire field for many years to come. (Look at "Go 2" versus "Perl 6", for example.)

On some level, everything is like that. Perl was a better awk/grep and was successful in replacing those tools. Then one day, someone decided to use it for CGI, and it got a huge popularity boost. But the people that decided to use it for CGI could have easily used something else, and it would have probably worked fine and Perl never would have been popular. It was just random chance that they picked Perl and made it popular. I feel like everything works that way. Javascript is popular for web UIs because someone decided to add it to a popular web browser. There is always some intrinsic randomness that plays a part in gaining mindshare. A bad tool with a good mindshare can do well, and a great tool with no boost can languish in obscurity. Perl benefited a lot from luck, but rolling the dice doesn't last forever.

If mod_perl had been designed more like mod_php Perl's web frameworks (Catalyst, HTML::Mason, Dancer & Mojolicious) might have had a chance of competing with PHP and Rails. PSGI could have helped Perl compete with Ruby/Rails but PSGI came too late.

Perl 6 is still very much a thing, but it's called Raku now (https://raku.org using the #rakulang tag on social media). It has a weekly blog post https://rakudoweekly.blog should you care to want to stay up-to-date!

To my knowledge, only 2 implementations were ever in development simultaneously: Niecza and Rakudo.

Niecza focused on producing a Perl 6 on the .Net infrastructure. Rakudo is based on NQP (Not Quite Perl) and initially ran on Parrot, then also added the JVM as a backend, and then later added MoarVM as a backend.

Niecza sadly stopped being developed as its main developer developer decided it wouldn't be able to come to a full release (around 2012/2013).

So to say that "about 3 implemetations of Perl6 in development simultaneously" is incorrect, and then to name MoarVM, JVM and Rakudo as the three is even more incorrect.

Perl 6 got renamed to Raku in 2019. Since the first release in December 2015, it has become about two orders of magnitude faster.

You can check out the Rakudo Weekly News should you wish to stay up-to-date on developments in the Rakudo world: https://rakudoweekly.blog

Of you're depending on a meme rather than making a valid point, I wonder what you are bringing to this particular table.

If it works, it works...

Reply on news.ycombinator.com