Future Directions for Parrot and Rakudo

Jonathan "Duke" Leto jonathan at leto.net
Fri Jun 21 20:23:50 UTC 2013


Howdy,

Sorry that I don't have time to give all the details, but here are some:

1) MoarVM is designed to be a virtual machine EXCLUSIVELY designed for NQP
2) Parrot will never be built on top of MoarVM
3) Parrot has amazing threads
4) NQP/Rakudo developers don't yet understand how to use Parrot
threads effectively (because we have horrible docs)
5) The future of Parrot has nothing to do with Rakudo Perl 6

Benabik, we are in agreement about everything, but there is so much
FUD, smoke and mirrors that it is hard to tell.

Duke

On Fri, Jun 21, 2013 at 11:03 AM, Brian Gernhardt
<benji at silverinsanity.com> wrote:
>
> On Jun 21, 2013, at 1:07 PM, "Jonathan \"Duke\" Leto" <jonathan at leto.net> wrote:
>
>> As soon as Rakudo Perl 6 works on MoarVM (the spiritual successor of
>> the m0 branch in parrot.git) [0], anything that Rakudo depends on that
>> is not deemed necessary by Parrot core developers will be removed.
>
> "Not deemed necessary"?  Necessary to what?  There are plenty of things in Parrot not really being used that stripping out the bits that are seems somewhat counter-productive.
>
> I would agree Parrot could use change.  Strip things down, don't be afraid to break things.
>
> If I had the time, I'd make a branch killing PIR and replacing it with a dumb assembly language and a bytecode generator a la ASM.  But I'd do that in parallel with Winxed and NQP branches that taught them the new system.  Building a system that nothing uses is interesting only from an academic standpoint (and I say that as an academic ;-).
>
> But I'd also want to do things like try to kill PMCs and rebuild them as 6model objects.  Pull _more_ from NQP, not strip things out from under it.  It's by far the most active thing on Parrot, and honestly I think points out a lot of the utility of the system.  The fact that the system is extensible enough to build something completely different is interesting.
>
> Heck, if MoarVM is a successor to M0, can we build Parrot on top of it?  I was never quite sure what would happen after an M0 interpreter was running, but maybe it would be interesting to try.
>
>
>
> Really, I guess my point is "we're removing things we don't need even if you're using them" is a very combative statement.  Especially when it's not clear what we actually need.  I'd say:
>
> 1) Don't be afraid to break things
> 2) Build interesting things
>
> but also
>
> 3) Talk about building first, instead of breaking first
> 4) Fix things we broke if we can
>
>
> ~~ benabik
>



-- 
Jonathan "Duke" Leto <jonathan at leto.net>
Leto Labs LLC http://letolabs.com
209.691.DUKE http://duke.leto.net
@dukeleto


More information about the parrot-dev mailing list