•Re: Re: Why I choose Perl

[merlyn]

Pick any five mature packages out of the CPAN, and five tasks they let you do for the cost of a download and a skim of the manpage.

Now ask your Python friend about how long it would either be to (a) handcode that in Python, or (b) find some package in their joke-of-a-non-CPAN to do the same thing.

"It's the code reuse, stupid." {grin}

This gets it out of the actual coding, and into the culture. Perl culture kicks hiney.

-- Randal L. Schwartz, Perl hacker
Be sure to read my standard disclaimer if this is a reply.

[cjf-II]

So what happens when Python and Java come out with their own archives? CJAN's on its way and a Python archive will be coming shortly as well (and they already have a good number of modules). Given the number of developers, I doubt they'll have much trouble finding contributors.

CPAN's advantage is going to be quickly reduced. If you want to prove Perl's superiority, why not focus on the actual language designs instead? Who knows, maybe the discussion will lead to improvements in all of the languages. Simply saying "Perl's better because it has more modules" leads me to believe you don't know Python well enough to comment on the actual language.

Preaching to the converted isn't much of a challenge either. If you really want a good debate on the subject, bring it up on a Python mailing list (in a nice non-flamebait way ;).

[Elian]

I think you very much overestimate the likelyhood that any of these archive projects will go anywhere. CPAN isn't new--it's been around for ages, and it arose from the ideas that were part of CTAN before it. This is not new stuff. Neither is it a new thought for the Python/Java/Ruby/VB/C/C++/Haskell/Scheme/Lisp/whatever communities to have their own. You'll note how many of those have materialized.

Building a CPAN network is dead-simple. It doesn't require much code, it doesn't require much talk, it doesn't require much planning, and it doesn't require much coordination. (Ask Jarkko) What it requires is a good chunk of sysadmin-ish effort on the part of a few people. You'll note how often that happens.

Archive networks have a chance of success when they're set up by someone with sysadmin skills, not programming skills. That's why so many attempts fail--programmers aren't sysadmins. Some very few programmers are also sysadmins, or sysadmins are programmers, but being a programmer confers zero (and sometimes negative) admin skills. Worse, most programmers wildly overestimate their admin skills, and often actively discourage those few who actually have them.

I'd love to see more CPAN-style networks set up. I won't believe one exists for a language until I actually see the damn thing in use. Until then it's fancy vapor with little hope of success.

[Aristotle]

See also the use.perl.org article with Jarkko's writeup on the topic.

Makeshifts last the longest.

[herveus]

Howdy!

I went to cjan.org. If there is a reasonable body of reusable code there, it seems to be well hidden. It appears that CJAN is still a concept, not an actuality.

I'm not trying to dismiss Java or Python; I know how to spell each. Beyond that, I can't make an informed comparison. I doknow that Perl and CPAN work for me.

CPAN appears to remain one of the strongest features of Perl. It also appears that "competing" languages are still a ways off from offering a similar, comprehensive solution for code sharing.

yours,
Michael

[cjf-II]

It appears that CJAN is still a concept, not an actuality.

By "CJAN's on its way" I meant the planning phase had begun. I should have made that more clear. See the mailing list for info on CJAN's progress :).