Larry Wall's Perl 6 Release Talk [video]
youtube.com
youtube.com
How does that work? is the moarvm bundled in the executable?
http://www.liz.nl/C-DayIsComingYAPCEU.pdf
So probably not "native" executables, but compiling and optimizing nonetheless.
Thanks Larry and co!
So please excuse my skepticism, because I'm still fascinated by Perl 6's ideas and potential. But is this another "Rakudo Star" where they just take whatever current mess they have and label it "release", or is Perl 6 really finally ready now?
That said, Rakudo of today is more usable than Rakudo of days gone by, though there are still issues (my personal pet peeve: parsing speed; hopefully, automatic precompilation will land in the not-too-distant future, which, would mitigate the problem).
The design documents at http://design.perl6.org/ are companion documents to the test suite, but where the two are in conflict, the test suite is authoritative.
As for changes: backward incompatible changes will result in a new language version.
I'd been naively holding out a hope that Perl 6 would improve things so I clicked on the Perl 6 for Perl 5 programmers link from perl6.org.
They added code block interpolation to strings:
#!/usr/bin/env perl6
my $number = 3;
say( "{ $number * 4 }" ); # prints 12 > perl -E 'my $i = 1; say "@{[ sin $i ]}";'
0.841470984807897
#!/usr/bin/env perl6
my $number = 3;
printf( "%s\n", $number * 4 ); # prints 12
I consider printf an acceptable form of string interpolation, nothing more.
I'd be interested to see what features there are that make the system slower in general, even when they aren't being used. I did a lot of work in my PhD on making odd features not affect the performance of Ruby code when they aren't being used.
Do you know what they are?
Do note that you will have to look closely and follow references to make sure you have a full picture of understanding, since at first glance many of them seem simple, but turn out to offer considerable depth.
Really, there hasn't been a terrible amount of new features in Perl 5 - and certainly not a new model for OO. Moose is great! And the constant release of CPAN modules is also wonderful, but the core of Perl 5 is full of dragons, which makes backporting features from Perl 6 hard to do in core - and perhaps a good argument can be made against doing that. Example: Moose is a very successful project, but thee given/when statement in the core of Perl was a failure.
CPAN is the language, perl is just the VM.
* multiple dispatch * the meta-object protocol
And the thing is, these features are being used all the time, because the core language is built out of them. If you want to see how a language which is not built on these things runs (on the same vm implementation, I might add), I suggest you try nqp :-).
I've come up with rather 'creative' workarounds[1] until a more proper solution arrives...
> I'd be interested to see what features there are that
> make the system slower in general, even when they
> aren't being used. I did a lot of work in my PhD on
> making odd features not affect the performance of
> Ruby code when they aren't being used.
You can type 'hi', ask folk to read your above comment (https://news.ycombinator.com/item?id=10565204 ), and then seen what unfolds when you chat for a few minutes. :)
They still have no idea how a fast dynamic language should be designed. To their advantage python does neither. Ruby does know, but doesn't get along with to well. Only PHP got their act together with 7.0.
BTW: Not Moose related - loop with native int: https://gist.github.com/mj41/20c70566564283151714
There's also the perl6 specific benchmark https://github.com/japhb/perl6-bench but I have no idea where they post their results to. It used to be a gist or IRC.
Note that to fully take advantage of it, the optimization layer above it that does type specialization also needs work (eg some optimizations were lost during the 'Great List Refactor', a semantic change that was set as one of the blockers for the 6.0 release).
Which languages with "similar features" are you comparing to Perl 6 here, and (given the second sentence) are the feature sets really similar?
All of the successful scripting languages have one: Perl 5 and earlier, Python, Ruby, Tcl, Lua, PHP, and even UNIX shell scripts.
Yet when it comes to Perl 6, we just don't see that.
The Perl 6 community has spun its wheels time and time and time again with half-baked interpreters written in an obscure language like Haskell, or written in Perl 6 itself using half-baked Perl 6 to 5 converters, or targeting the .NET CLR or JVM, or targeting very obscure and limited VMs like Parrot or MoarVM.
It's no wonder that Perl 6 hasn't really gone anywhere after 15 years. They've done everything but the one thing they should be doing if they want success: implementing a traditional interpreter using C that runs just about everywhere!
For 99% users "Perl 6" will mean just "Rakudo + MoarVM" which seems to be exactly what you're asking for.
Note that MoarVM is nothing like Parrot. Parrot was a failed project which intended to bring universal VM for all dynamic languages. Obviously, too wide project scope resulted in broken and unmaintainable implementation.
MoarVM is made just for Perl 6 and intends to avoid repeating its predecessor's mistakes.
Rakudo and MoarVM clearly aren't what we, as potential Perl 6 developers and users, want.
They look far too similar to the many other similar Perl 6 failures we've seen.
How are we to trust that Rakudo won't go the way of Pugs, or that MoarVM won't go the way of Parrot, leaving us in a lurch?
Even if they do remain viable, we don't want to have to deal with separate VMs, regardless of whether they're specific to Perl 6 or not.
We don't want to deal with the bugs and compatibility issues that invariably pop up when dealing with separate compilers and VMs.
Like I said, we want the Perl 5, Ruby, Python, Lua, and Tcl type of experience.
We want there to be one consistent implementation that's widely used and trusted.
We want something proven and familiar.
Unfortunately, we aren't getting that with Rakudo and MoarVM, nor from the many other unsuccessful attempts in the past.
Rakudo + MoarVM is going to be "it" going forward I would think. I guess we'll wait and see. That doesn't mean that other compilers won't come along because that is one of the benefits of Perl 6. All the languages you mention have different versions out there. The only one that doesn't is Perl 5 and that's because only Perl can parse Perl. That's going away with Perl 6 and that's good. Rakudo+MoarVM is great. It still leaves the door open for genius to shine.
Rakudo + MoarVM is exactly what potential Perl 6 developers want. Everything else is just gravy.
I have no idea if Perl 6 will "take off". I hope it gains traction. Watching Larry Wall got me interested enough to install it and start learning it anyway.
There's CPython, the main Python interpreter. It's quite traditional in architecture, and is written in a very portable subset of C.
It's the Python implementation that pretty much all Python developers and users happily and successfully use.
Then there are the experimental or niche Python implementations. They augment very specific and isolated use cases, instead of splintering and delaying the adoption of the language like the many failed partial Perl 6 implementations did.
It should be evident that the Python approach is better than the Perl approach. We've been able to use Python 3 in production for many years now, while Perl 6 still isn't usable.
I'm historian of Perl 6, a bit :-). http://www.slideshare.net/michaljurosz1/perl-family-15-years...
Open source, volunteers, real life, ecosystem, not easy to implement so feature rich language, ... this wasn't planned.
You've got MoarVM, which is just like CPython But if you want a different backend - great! Don't have to wait for MoarVM - the folks at MoarVM didn't wait for Parrot to solve their problems (or they'd still be waiting). For example, use JVM, just like Jython.
This is a stark difference to the situation of Perl 5, where there is only one interpreter, and there's little, if any chance of another. There's no Perl 5 language specification (as there is in Perl 6), except the Perl 5 interpreter. Only Perl 5 can parse Perl 5.
Perl 6 got so many things right, it's phenomenal.
Getting things right takes time, and it only happened because so many things went wrong.
This historical revisionism continues to be tiresome.
Some people will tell you to trust them. I, on the other hand, base my expectations on the history of the project and the actions of the people involved. I don't believe you can turn 15 years of disappointment and underperforming into something usable based on an arbitrary deadline enforced by yet another marketing announcement.
One can hope that development focus will then switch from semantics to things like stability, deployment and performance, ie the things you strongly advocated for years ago.
I guess we'll see if it's going to happen that way...
Given the reception to the past year's worth of discussion around "Get ready to party", it seems irresponsible to me to continue with these semantic distinctions, as they're lost on the intended audience.
Why not? A core implementation with its own VM (MRI+YARV for Ruby from 1.9 on, for instance) seems to be pretty common for languages.
> Like I said, we want the Perl 5, Ruby, Python, Lua, and Tcl type of experience.
But that's exactly what Rakudo + MoarVM offers. For the Ruby comparison, Rakudo is analogous to MRI, MoarVM is analogous to YARV.
> We want there to be one consistent implementation that's widely used and trusted.
And why can't Rakudo be that implementation?
> We want something proven and familiar.
Well, clearly, it won't be proven till its been around for a while in production use, and won't be familiar till you've used it for a while. But I don't see any reason that Rakudo (which is used for some things exposed to the public now) can't be the thing that ends up "proven and familiar"?
> Unfortunately, we aren't getting that with Rakudo and MoarVM
In what concrete way is that true?
Because the author thought it to be fun. You know, hackers and all.
> or written in Perl 6 itself using half-baked Perl 6 to 5 converters
Oh, I don't think that ever happened.
You may be thinking of NQP? Which is still a Thing, and it's what Perl 6 is written in, which is amazing.
There was also the initial idea that Perl 6 could run Perl 5, but this is proving impossible. This is done now using the Inline::* modules, which will use say, the Perl 5 interpreter. [0]
So, there's one for Perl5 to allow for compatibility with Perl 5 modules, but also for Python[1], and most likely for x language. That's awesome!
Of course, the inverse is true, too. You can call up Perl 6, from Perl 5![2]
> or targeting the .NET CLR or JVM, or targeting very > obscure and limited VMs like Parrot or MoarVM.
Targeting .NET or JVM is also a good idea! Many languages do it. The general idea of Parrot shares a lot of interesting ideas with .NET - I almost want to say Parrot precedes .NET. The April Fools joke certainly did. If it didn't, Parrot was certainly inspired by .NET - or it was just an idea that was in the air at the time. Good idea!
[0] https://github.com/niner/Inline-Perl5 [1] https://github.com/niner/Inline-Python/ [2] https://metacpan.org/pod/Inline::Perl6
Oh, I don't think that ever happened.
That might be a reference to viv, which was used to implement STD (cf http://perl6.org/compilers/std-viv ).
There was also the initial idea that Perl 6 could run Perl 5, but this is proving impossible.
Personally, if Rakudo ever starts outperforming Perl5 for practical workloads, I believe that idea might eventually be revisited.