Hi,
I think featurewise Rakudo is now at a point where it could already be
use for some serious work. Surely many things are missing, but (for me)
the two most important things - good OOP support and types - are already
in. And the syntax is just lovely :) (I think I have a syntax-fetish... :))
However, performance is an issue. I would not mind running into bugs,
writing some extra code to work around missing stuff, etc., but right
now it is just hard to find any projects (for me - YMMV) where
performance would not be a blocker. Right now the only thing I can use
Perl6 for is to learn Perl6, and that's a problem, because when I sit in
front of a computer, learning is the only thing I do not get paid for :)
For a long time, people I talked about Perl6 all feared that it would
never be ready. Now concern shifted to "will it ever be fast enough?"
And I think it's not really about Rakudo, but about Parrot.
Unfortunately I am not a Rakudo, neither Parrot hacker, but as I
understand, there are features in Rakudo which are implemented using
only a few lines of code. You cannot gain 100x speed difference
optimizing, say 7 lines, so I must assume it really is Parrot which
needs a lot of love, and there aren't too many people who can/could hack
on Parrot. That sometimes feels scary.
I very much agree with Patrick: an order-of-magnitude speed difference
compared to Perl5 is kind of the point where many will just stop caring
about performance and start using Rakudo/Perl6. Actually I expect a
significant increase in the number of new Perl6ers at around < 100x
slower. (That, and the 10 "most important" Perl5 CPAN modules ported to
Perl6 :))
I think it would be nice to have some sort of a performance-tracking
page, with just some very basic Perl5 / Perl6 code and very rough
measurements. If noone will do that, I will :)
- Fagzal
