It is the ecosystem

[zby]

There is an excellent slideshow entitled Perl Myths by Tim Bunce of DBI fame, busting the long lived Perl misconceptions. I would post it to the News section of this site - but I have a couple of comments on it.

I'll start that the numbers showing CPAN expotential growth should not be treated as Perl specialty - it is Open Source as a whole that has this property. But still no other language has so much infrastructure built for making the openly available code useful. CPAN was the first of a kind repository and it is also the first one to reach such a massive scale forcing us to find ways to keep it useful under this expotential growth. And this is something that will not be easily copied by other languages - because it is an ecosystem - a complex conglomerate of software tools, social practices and supporting organisations. The testing culture with supporting modules, dependency scanning , Perl::Critic, mailing lists, IRC chatrooms, TPF and our own site are all elements of it. This is where the language maturity really shows off. And there are new experiements mashrooming all the time - like Annotated CPAN documentation and CPAN::Forum (which is, by the way, waiting for some Google Summer of Code applicants to write a mailing list gateway and other improvements).

And in the face of expotential growth of all Open Source software Perl is once again in the avant garde treading the new ways of maintaining a massive ecosystem of interdependend code libraries.

Update: There are other versions of that talk (including a video) linked from Tims blog
Related:Have you heard about recent startups using Perl?, What can bring the excitement back to Perl?
Update: Divine Invention: An interview with Larry Wall: "Most other computer languages are pretty sterile. They're about technology. The difference with Perl is that I decided to create the culture at the same time as I was creating the language."

[sundialsvc4]

Perl is a language that somehow escaped my attention for many years. I wish it hadn't, and I'm not sure quite how it did. There's no doubt that the language, quirky though it is, can “move the freight” in a lot of practical situations. I know I wasted a lot of billable time for not knowing it.

The CPAN library is great, and "here's a rose"   @>-+---   to all the folks who work so hard on it. Perl would not be what it is without you, and we all know it.

I'd observe that Perl did a lot of things to naturally attract this sort of attention to itself, because it was a reasonably well-implemented string-processing language at a time and at a place when there was a lot of demand for that sort of thing. Aside from the pedantic purposes that Larry Wall originally built and used it for, which are themselves important and legitimate, Perl is a real “Swiss army knife” for web-sites. It was a purpose-built tool that was, all things considered, well built and re-built. People picked it up and started using it, and “results happened.”

You can really tell when a programming tool was designed by one person, or by a small group of persons who actually need to use the tool, vs. a committee of academics who want to make a fashion statement. The latter group builds a language with a 58-megabyte runtime (actually three of them, all different) that still can't print reliably on Linux and still can't get out of its own way when doing simple GUI-applications on any system. (I am sure by now that you know exactly what (de-)caffeinated language I am talking about.)   :-D

[brian_d_foy]

Actually, CPAN was not the first of its kind. CTAN, the Comprehensive TeX Archive Network beat CPAN by a couple years, and it's not an accident that they have a similar name. I find CTAN just as useful as CPAN, and for the same reasons.

It's extremely easy for any other language to copy this. All they have to do is follow Jarkko's advice to make something simple that other people can build on. Perl has the ecosystem because Jarkko and Andreas made a good foundation and then got out of people's way. Today's kids have too much technology, so they think a simple FTP archive is stupid. They try to build overly-complex systems which they then lose interest in.

[zby]

Thanks for the historical corrections. To copy the archive technology is easy, to copy the community customs around it (and by this I mean for example the pressure on testing, the module naming conventions etc) is not so - this is my main point.