This is complete vaporware. The constraints and software development link above is just annoying. There is no excuse for an 8 year long dev effort with no milestones or end in sight. We could have had a nice language by now with some simple features like parameter lists...but we're stuck with Perl 5 circa 1996 because some people got big useless ideas into their heads and hijacked what was once a useful technology and turned it into an obsolete pile of code about as valueable as awk or sed. The article should really begin to change, from a "whats new and cool in [Perl 6]" to "why [Perl 6] was such an utter failure". the [Perl 6] team has failed us all.
— Justforasecond, Personal Comments on the Perl 6 discussion page on Wikipedia
If my economics are wrong, show me. If my math is wrong, show me. If you know the secret to software development, show me. Here's zero dollars and the source code. Fix everything you can about Perl 5. See you in two years.
Here's 1084 patches I've made or applied to Parrot in the past three years. Not coincidentally, I've had actual Perl 6 code running in public for about the same amount of time. You can find it if you search the web. Of course, you can find Perl 6 and Parrot milestones if you search the web too.
P.S. Perl 5.10, circa 2007 is quite decent. Try upgrading to software released this millennium.
but we're stuck with Perl 5 circa 1996
Anyone who thinks that Perl 2008 is just remotely like Perl 1996 should have his perception ability examined.
From the 1996 Perl thinking "we don't need no steenking objects" over the new milenium's "it's my code, I will obfuscate it" we've arrived at a very healthy development thinking, with Perl running on all cylinders full throttle.
When you develop a taste for readable and maintainable code, your choice of tools and libraries will change too. There are more CPAN modules and systems that work in a smooth and straightforward way now. Catalyst, for example, lets you divide your code in easy to maintain modules. I think everybody as a shortlist of sane modules from CPAN they rely on to get the job done.
I know, Perl can be frustrating old hat with its poor parameter passing and strange (de)referencing syntax. But I much prefer to code in a language with some strange warts than in a language I have to fight all the time to make it do what I want.
Re:Perl 6
Stevan on 2008-04-19T14:21:55
Would we have Moose if Perl had a strict object system?See Joose for an example of Moose-ness in Javascript, although to be fair the JS object system is not really so strict, but certainly not as loose as Perl.
And if you want to see something really scary, take a look at this Java version of the Perl 6 metamodel which Moose was based on. If ever there was a strict object system, Java would be it. Of course the surface syntax of a Java-Moose would be intensely verbose and completely unusable unless you basically implemented some kind of pre-processing syntax-extensions of some kind.
But your correct, very few languages that I know would allow me to write Moose and integrate it to tightly into the existing code/culture the way Perl 5 did (only LISP/Scheme come to mind). But then again, there are not many others languages in which I would have needed to write Moose, cause they have decent built in OO support already. It is really a double edged sword.
-StevanRe:Perl 6
chromatic on 2008-04-19T17:06:22
As I see it, he would be happy with a Perl 5 with better parameter passing and smaller things like this. An he thinks that Perl 6 has prevented this.History doesn't support experiments; there's no way of knowing if that's true. Also I suspect that if you found ten people dissatisfied with some aspect of Perl 5 and asked them to design one feature apiece to fix Perl 5, they'd all choose something different. (History supports that assertion. We call it the Perl 6 RFC process. It didn't work. I'm not saying that it could never work, but it didn't work.)
It's true, it seems that development of new features Perl 5 was more or less stopped for some time, was this because of waiting for Perl 6? Or was it because nobody could maintain the Perl 5 sources.I'm sure there was some effect, but again this is very difficult to measure. If you look at the people who worked on both Perl 5 and various pieces of Perl 6, there's some overlap (Audrey, Nicholas Clark, myself) but not a huge amount. Different people are doing different things.
Let's try to do just that.
Here's zero dollars and the source code. Fix everything you can about Perl 5. See you in two years.
I disagree with your economics. While the sums spent on Perl 6 development were certainly not vast, it may actually be one of the better funded open source language developments out there, certainly compared with Perl 5 during the late 20th century period in which I followed it closely. Are Ruby, Python, Lua, and, for that matter, Perl 5.x, really better funded than Perl 6?
Here's 1084 patches I've made or applied to Parrot in the past three years.
Here's one secret to software development, especially open source software development: It does not count as "coding" until you ship it in an actual release >= 1.0. Until then, it's just intellectual masturbation. The problem with Perl 6 has never, as far as I know, been a shortage of people with ideas, or willingness to write code to back up those ideas. There seems to be, however, a dire shortage of people willing to commit to shipping and to building a consensus to focus on shipping, to the exclusion of any new neat idea that may come along.
Until Perl 6 has a demonstrable use to the public, it's entirely pointless to complain about economics. Perl 6 has received a fairly generous advance of money and trust from the community, as open source projects go - including people willing to buy a book about it 5 years ago. It's up to Perl 6 now to earn back that money and trust by shipping and being of actual use. Once people can use it and don't want to live without it, asking for more money is entirely reasonable.
Not coincidentally, I've had actual Perl 6 code running in public for about the same amount of time.
Is it deployed? Is it expected to keep running the next time Parrot / Perl 6 are revised?
Of course, you can find Perl 6 and Parrot milestones if you search the web too.
The two milestones I've heard personally, from people high up enough in the project to count, were "Shipping code in 18 months", stated in 2000, and "Shipping code in 18 months", stated in 2004. Admittedly, it's been several years since I've actually cared about when Perl 6 would ship, other than out of a morbid sense of curiosity.
P.S. Perl 5.10, circa 2007 is quite decent.
And so was Perl 4.019, circa 1991, and most releases in between. Perl was steadily moving forward until 2000, and 5.x again seems to be steadily moving forward now. Is the Perl 6 project helping Perl as a language? I don't think so. It absorbed a lot of mindshare and momentum in the early 2000s, it cast FUD on the continued viability of Perl 5 (and has vastly benefited Python, and to some extent Ruby, in the process), and it arguably still siphons off funding that would be better spent on moving Perl 5 along. The result is a steady shrinkage of the Perl ecosystem, even though it's still formidable.
That said, y'all of course have every right to do what you want with Perl 6. Just don't be surprised if your efforts to figure out how many continuations can dance on the head of a sigil are met with more derision than awe these days.
Re: Those with Loaded Mouths, and Those Who Code
chromatic on 2008-05-08T00:19:17
Are Ruby, Python, Lua, and, for that matter, Perl 5.x, really better funded than Perl 6?Charles Nutter, Ola Bini, John Lam, and at least one more person I'm forgetting at the moment are all funded Ruby implementation developers.
Guido van Rossum is a funded Python developer. I don't know whether to count Alex Martelli. Jim Hugunin is a funded Python developer.
I don't know who's funded to work on Lua. I thought it was a research project from Roberto's university, but I haven't confirmed that just now.
To find funded Perl 5 developers, you have to go to ActiveState, and the most active contributor I've seen on p5p from AS is Jan Dubois.
Perl 6 has Jonathan, one day a week for at least three months, Patrick for three months, and a double-handful of milestone payments for which I don't even want to do the dollars-per-hour math.
Even Haskell has more full-time funded developers than Perl 5 and Perl 6 combined. Even Tcl has more full-time funded developers than Perl 5 and Perl 6 combined.
Here's one secret to software development, especially open source software development: It does not count as "coding" until you ship it in an actual release >= 1.0.I invite you to stop using all free or open source software with a version number less than 1.0, even if it meets all of your needs (as if version numbers had any connection with reality).
The problem with Perl 6 has never, as far as I know, been a shortage of people with ideas, or willingness to write code to back up those ideas.This conjecture is also wrong.
Re: Those with Loaded Mouths, and Those Who Code
murr on 2008-05-08T17:43:12
Charles Nutter, Ola Bini, John Lam, and at least one more person I'm forgetting at the moment are all funded Ruby implementation developers.My impression is that Nutter & Bini are on JRuby and Lam is on IronRuby, so I'm not sure they should count for funding of the mainline interpreter.
Guido van Rossum is a funded Python developer.
Full time? My impression was that he had other duties at Google as well.
To find funded Perl 5 developers, you have to go to ActiveState
Who, I believe, tend to fund development as a side effect of fixing bugs for which they have customer support contracts (which again is open source economics: projects that ship and are useful can generate some revenue to move forward said projects).
Perl 6 has Jonathan, one day a week for at least three months, Patrick for three months, and a double-handful of milestone payments for which I don't even want to do the dollars-per-hour math.
... plus, judging from the recent Perl 6 design minutes, some Google Summer of Code and ASF grants. I'm not arguing that this adds up to huge amounts, but if you look at it over the almost 8 years of Perl 6 development, rather than on a monthly cash-flow basis, it is not all that insubstantial either. I invite you to stop using all free or open source software with a version number less than 1.0, even if it meets all of your needs
That would affect, roughly, OpenSSL and apt.
(as if version numbers had any connection with reality).
There are, of course, exceptions, but in general, 0.x means "Not ready for prime time". Are you arguing that Parrot or Perl 6 ARE ready for prime time?
Re: Those with Loaded Mouths, and Those Who Code
chromatic on 2008-05-08T18:01:22
... if you look at it over the almost 8 years of Perl 6 development, rather than on a monthly cash-flow basis, it is not all that insubstantial either. Let's take the $150,000 paid to Larry, Damian, and Dan, the $70,000 from NL.net, the $10,000 from the Mozilla Foundation, $3000 from Vienna.pm, and $20,000 from various Summer of Code projects. That's $255,000 over eight years for a language design, at least one modern virtual machine, and one implementation. That's just under $32,000 per year.
I invite you to tell Matz he can get Ruby 2.0 for $32,000 a year or Guido he can get Python 3000 for $32,000 a year, or Anders he can get
.NET 3.0 for $32,000 a year, or Gosling he can get Java 7 for $32,000 a year, or Rasmus he can get PHP 6 for $32,000 a year. Which do you want, fast or good? You can't not pick cheap.
Are you arguing that Parrot or Perl 6 ARE ready for prime time?You've changed your original argument, so I have no interest in continuing.
Re: Those with Loaded Mouths, and Those Who Code
murr on 2008-05-08T18:49:27
... if you look at it over the almost 8 years of Perl 6 development, rather than on a monthly cash-flow basis, it is not all that insubstantial either.Let's take the $150,000 paid to Larry, Damian, and Dan, the $70,000 from NL.net, the $10,000 from the Mozilla Foundation, $3000 from Vienna.pm, and $20,000 from various Summer of Code projects. That's $255,000 over eight years for a language design, at least one modern virtual machine, and one implementation. That's just under $32,000 per year.
My point is that it's irrelevant how much it is PER YEAR. Also, if I read the Perl foundation tax returns correctly, Larry, Damian, and Dan got more like $265,000, and there were some other nontrivial grants from the Perl foundation, so the total (assuming that your other numbers are correct and complete) would be more like $400,000.
I invite you to tell Matz he can get Ruby 2.0 for $32,000 a year or Guido he can get Python 3000 for $32,000 a year
Offer either of them $400K for those milestones, and I don't think they'd dismiss the offer as unrealistic.
Are you arguing that Parrot or Perl 6 ARE ready for prime time?You've changed your original argument, so I have no interest in continuing.
I've changed the wording of my argument, but not its thrust, which is that work on a project only starts counting once that project is reasonably stable, usable, and feature complete.
Are you arguing that Parrot or Perl 6 ARE those things?
