|
Re^2: What's wrong with Perl 6?
by TimToady (Parson) on May 11, 2007 at 00:41 UTC
|
But I clearly remember the first stabbing pain that told me it was going to be something different than first advertised.
I confess I don't quite understand this. Than first advertised by whom? As far as I'm concerned, the first advertisement for Perl 6 was simply that we were going to look at Perl as a community and fix everything that needed fixing, and that it would certainly include making Perl powerful enough that we could write its parser in Perl itself. If things have changed from that, it's perhaps a sense of reality about how many things there were to fix. I expected maybe 20 RFCs, and we got 361. Our eyes were opened to the fact that pretty much everyone had tunnel vision about how to change Perl for the better, and that the sum of the community's pain was much greater than any individual piece of it. To me, we are still precisely on the originally announced target of having a community rewrite of Perl. We just didn't understand the scale of what that meant.
We also didn't understand the full ramifications of what it means to have a Perl 6 compiler written in Perl 6. Among other things, it means you don't really have to care whether the backend is using Haskell or Parrot or the latest and greatest VMs from MS and Sun. The current Haskell implementation of pugs is a nice prototype, but in the long run, Haskell is just another engine to run on.
As mentioned elsewhere, fglock has already bootstrapped a mini-Perl 6 in itself. And we're a goodly part of the way to having a full Perl 6 parser written in full-up Perl 6. See the standard grammar, which we can currently parse, but not quite run yet. But it's getting there. (Would you trust a precompiled compiler written in Perl 6?) Once we get to the point of being able to run it (on any of the engines), things will converge rather rapidly from there on.
| [reply] |
|
|
I confess I don't quite understand this. Than first advertised by whom? Somehow this reminds me of TV advertising. :-) What I remember (subjectively and from a perspective of a wannabe on the sidelines) was that Perl6 was going to be this great revolution in interpreted languages, where everything was going to be built on the Parrot ( a universal assembly language), and this would allow for Python, Perl5, Ruby (and whoever else wanted in) code to be seamlessly inlined into Perl6 code. Now if my impressions were wrong, you can point that out, but I think many others were given the same impression. Maybe it was just part of the Parrot hype.... and Perl6 shouldn't be blamed for it....but it all has become confusing to me. Can you estimate how long it will be before Perl6 can be compiled from C, like Perl5?
| [reply] |
|
|
| [reply] |
|
|
|
|
|
|
|
Among other things, it means you don't really have to care whether the backend is using Haskell or Parrot or the latest and greatest VMs from MS and Sun. The current Haskell implementation of pugs is a nice prototype, but in the long run, Haskell is just another engine to run on.
Well, sometimes if fever gets me, I'm dreaming of a Perl 6 implementation in FORTH, but that's perhaps just my delirating style... ah, a perl chip! That would be IT! :-)
--shmem
_($_=" "x(1<<5)."?\n".q·/)Oo. G°\ /
/\_¯/(q /
---------------------------- \__(m.====·.(_("always off the crowd"))."·
");sub _{s./.($e="'Itrs `mnsgdq Gdbj O`qkdq")=~y/"-y/#-z/;$e.e && print}
| [reply] |
Re^2: What's wrong with Perl 6?
by Fletch (Bishop) on May 10, 2007 at 19:10 UTC
|
I don't hate Haskell, but Perl5 compiles directly from C, and I'll never trust precompiled compilers.
So wait . . . your C compiler sprang forth fully formed from the brow of whom exactly, Kernighan or Ritche? :)
| [reply] |
|
|
Sure you can make that comparison, but there is a vast army of programmers watching that gcc compiler, and I can trust that in those numbers, someone will start complaining if anything looks "fishy" in the code. The same isn't true with Haskell. The gcc I use can be bootstrapped from itself, although most people do trust the prebuilt Glib and compiler that comes with their distros. When I get time and energy, I usually try to recompile them for myself, just to see the differences. So this all brings up the ultimate question of whether people will trust Perl6 ...... if there is this abstract Haskell layer in there. I confess, I think I've just been given a negative feeling toward the word Haskell
after watching Leave It To Beaver reruns all these years.
| [reply] |
|
|
Hey, we're just reciprocating. If Haskell (GHC) can use Perl to bootstrap itself, we can use Haskell to bootstrap Perl 6. :-)
Seriously though, the plan for as long as I've known has been to write an implementation of Perl6 that is enough to start writing the Perl 6 compiler in Perl 6. So, the "haskell layer" is just part of the bootstrapping phase and most people won't be aware of it because it'll be long gone by the time they get to running actual Perl 6 code.
| [reply] |