[chromatic]

People said we'd never release Parrot 1.0 either (people said that as recently as November) -- but why speculate when you can measure?

We released Parrot 1.0 in March. We released Parrot 1.1 in April. We released Parrot 1.2 in May. We'll release Parrot 2.0 in January 2010 and Parrot 3.0 in January 2011.

Keep that in mind as you look at the daily Rakudo status reports. As of last week, Rakudo passed 68% of the current spectests. The passing test velocity has only increased this year.

[rjurney]

Calling it Parrot 1.0 doesn't mean its stable. You can call it 10.0, doesn't mean people are gonna switch.

I'm not as well informed as you, but I'm just really skeptical about a 10 year old project shipping 1.0. I'm pissed off that Perl has died in the meanwhile. Something went terribly wrong and the language was mis-managed.

[chromatic]

What else would we call Parrot 1.0? We believe that Parrot 1.0 represents a stable platform on which people can start to build compilers.

It's not a finished platform (which is why we continue to produce new releases), but we have a documented and well-understood deprecation policy.

We'll add new features, but we believe that the current set of features in Parrot 1.0 is sufficient to build a workable language.

(As for the question of "Is Perl dead?", the rate of uploads to the CPAN certainly disagrees. That's a measurable data point. I won't address the question of Perl 5's release policy and backwards compatibility concerns here; I've discussed them at length elsewhere.)

[rjurney]

Why didn't Perl 6 just use the JVM? Why did you decide to build Parrot?

As to CPAN uploads: Nobody I know under 30 in Atlanta knows Perl. I'm not exaggerating. Thats the important data point to me.

[chromatic]

That's a very silly data point for several reasons.

First, the age of 30 is arbitrary, unless there's some reason that only the experiences of people under 30 matters. Perhaps everyone over 30 stops coding, or dies, or fails to create anything interesting -- but you haven't demonstrated that.

Second, the choice of Atlanta is arbitrary. Is Atlanta a sufficient statistical representative of all of the locations of programmers in the world?

Third, your choice of "people I know" is arbitrary. Can you demonstrate that you know a representative sample of available programmers in the Atlanta area? (What happens if you expand the definition of "programmers" to include "people who occasionally write a program"?)

Fourth, your experience doesn't compare. An anecdote is not a piece of data.

The rate of change on the CPAN is not the sole determinant of Perl's viability, but there is a single, well-understood place to share reusable code with other Perl programmers. It's measurable data, and it's normative for certain types of Perl usage.

What you provided isn't data. It's just noise.

[rjurney]

http://www.tiobe.com/index.php/content/paperinfo/tpci/index....

What about that noise? This is pretty typical Perl thinking: pretend everything is ok, nothing to see here. Lots of CPAN commits, we have one metric to cling to. Everything is ok.

[chromatic]

TIOBE's flaws are well-documented throughout the Internet (look out Ruby -- Delphi, RPG, and Logo are hot on your heels!). In particular, no third party has ever been able to duplicate their results, whereas anyone with access to any of the hundreds of CPAN mirrors can verify those numbers.

[draegtun]

Here is another metric we can cling onto: http://www.blackducksoftware.com/news/news/2009-01-21 ;-)

Generally I take everything I read with a large pinch of salt unless all the facts are presented and can be proven so don't believe either of these are fact especially the TIOBE one!

[Zaak]

> Why didn't Perl 6 just use the JVM? Why did you decide to build Parrot?

http://www.perlmonks.org/?node_id=142797

[brunov]

I'm 24 and I code in Perl 5. Do I count in your poll, or do I have to live in Atlanta?

[rjurney]

If I said nobody whatsoever under 30 codes Perl, you would count in my poll. I said nobody I know. I realize there are some of you, but you must know that there aren't many of you, right?

[jrockway]

Most people I know that code Perl are in the 25-30 range, and I consider myself pretty widely connected in the community. Why do you care about age, anyway... it's a programming language, not a boy band.

[brunov]

Ok, here's your premise: Perl is dying. And here's your argument: I don't know anyone below thirty in Atlanta that uses it.

Don't you actually see where your logic slipped? Sure, CPAN having doubled their upload rates last year doesn't mean anything, but the few young acquaintances of your personal social circle really add up to your statistically proved statement.