Perl is not Dead, it is a Dead End | Hacker News Reader

Perl is not Dead, it is a Dead End

speakerdeck.com

38 pointsbrunov13 comments

This is worth reading, in spite of it's inflammatory title, as it's really about a new project: moe ( https://github.com/stevan/moe ) an "ultra-modern perl 5" written in Scala.

Some of the problems that the author mentions in Perl5 are actual problems in Scala, for example the compiler and interpreter latency or start up time, the complexity of the language, the philosophy TIMTOWTDI and the bigger one is that other languages are getting better and better. Why just not to translate the philosophy of Perl to Scala, that is write a book translating the most common idioms of Perl to Scala. Good programmers could learn the Perl-Scala way in days, the real thing to realize is that today there is no single solution, you must master several programming languages, no one size fit for all problems.

I mostly agree and believe that before another abomination loosely based on Perl is written in Scala, people should learn Scala. However, Scala's syntax is also quite complicated, Scala code is often hard to read (far too much TIMTOWTDI there as well), so personally I favour Go as a modern alternative to Perl.

The author doesn't make the case all that strongly. Despite the quantity of words there isn't a lot of content backing them up. Mostly "fun" quotes an in-joke references I don't get, which he seems to bemoan early in the presentation but lapses into about halfway through.

Perl 5 never struck me as something one would set out to acquire from scratch. Perl 6 does, but there are fresher faces and I wouldn't predict it becoming a major force anytime soon.

great slides; I read the whole thing. This is a very interesting project, too. I left perl around the time of Moose because I was desperate for a sane OO system. I'd love to see this succeed.

If "Perl is a dead end" (his words, not mine), then why not switch to another similar language like Python or Ruby?

I guess I don't understand why this guy wants to create yet another programming language. He spends his entire slide deck complaining about how bad Perl is, without ever discussing the obvious possibility of... just using another language. Is there anything good about Perl 5 that other languages haven't already copied? If there is, what is it? I am not trying to be facetious here-- just curious.

> If "Perl is a dead end" (his words, not mine), then why not switch to another similar language like Python or Ruby?

Because a lot of the points given also apply to Python & Ruby.

>I guess I don't understand why this guy wants to create yet another programming language.

He isn't. It looks like he's creating an improved / cleaned-up perl5 running on a new VM. Think of it as perl6 if the perl6 moniker hadn't already been taken!

EDIT - Hopefully clearer analogy... the project is a mixture of python2 -> python3 (language cleanup) + MRI -> YARV (new compiler & VM).

> Is there anything good about Perl 5 that other languages haven't already copied? If there is, what is it?

The author of the talk (Stevan Little) is the creator of Moose (http://moose.perl.org). I think what he maybe striving for is a Moose implemented on a modern future proofed VM (full threading/multiprocessing, etc).

The question you may want to ask is why? By sounds of this talk it appears that his p5-mop proposal/work (https://github.com/stevan/p5-mop), which was originally mooted for perl 5.16, as probably been blocked by the Perl core team (p5p). It's a shame if that is what's happened :(

> The question you may want to ask is why? By sounds of this talk it appears that his p5-mop proposal/work (https://github.com/stevan/p5-mop), which was originally mooted for perl 5.16, as probably been blocked by the Perl core team (p5p). It's a shame if that is what's happened :(

That's what he hints at in one of the slides. I'd like to know what exactly happened though, I was interested in that project.

Same here.. I'd definitely like to buy Stevan a beer to find out whats happened.

In meantime I raise my glass (again [1]) for all his great work on Moose and hope that p5-mop does make it into Perl core and/or he manages to make Moe a success.

[1] - http://stevan-little.blogspot.co.uk/2010/03/moose-100-is-rel...

Because a lot of the points given also apply to Python & Ruby.

Like what, exactly? He even has a slide which says this about Node.js: "there is no problem." If there is no problem, then what is he trying to solve by creating a new language? I am confused.

Moose is cool and all, but it basically just brings Perl up to the same level of functionality that other languages like Python and Ruby have right out of the box. Using the JVM seems like a particularly unfortunate choice since it has such a long start-up time and traditionally Perl scripts were for quick and dirty jobs.

>Like what, exactly?...

Threading / Multicore processing. This is the 800 pound gorilla sitting in all programming language camps.

>Moose is cool and all, but it basically just brings Perl up to the same level of functionality that other languages like Python and Ruby

Nope... Moose brings Perl up to the CLOS (Lisp) level of OO which is beyond what you get with Python & Ruby.

Moose is cool and all, but it basically just brings Perl up to the same level of functionality that other languages like Python and Ruby have right out of the box.

Python and Ruby have parametric roles/traits? Typed attributes with coercions?

Slides talk, numbers walk.

It's not fat Perl hackers and CGI weenies anymore, but rather Python dev-ops wahoos and Rails coders.

There are all kinds of wonderful toys old and new that I'd love to be using. I just don't have a practical reason to call upon them.

Reply on news.ycombinator.com