Never end? More like never deliver.
http://rakudo.org/downloads/star/
Same as firefox... or chrome... or any other project on a continuous release cycle.
The project that led to Firefox was started in 1993. It was called Netscape. (Let's say this is analogous with Perl 5.)
In 1998 the project was reinvented as Mozilla. (Let's say this is analogous with Perl 6.)
For much of the 7 years between early 1998 and late 2004 the Mozilla project (analogous with Perl 6) was lambasted by the masses for, first, failing to release anything, and then, when they did, releasing slow, buggy, bloated, irrelevant software. By 2003 Mozilla was being ignored by almost everyone.
What saved Mozilla was that some smart engineers refused to bow to, on the one hand, public disinterest, and, on the other, internal pressure to focus on improving Mozilla 1.0. Instead they did yet another rewrite of several key parts, and then in early 2003 those still interested got behind the rewrite and the outcome was Firefox 1.0 in late 2004.
The parallels with the Rakudo team's efforts are clear and I see scope for an analogous outcome.
Beyond that, your attitude is easy to compare to a 2000s "SSJS will never succeed".
I'm not certain that yet another mostly-incomplete virtual machine will really help the Perl 6 community much. They've already got that with Parrot.
The whole emphasis on targeting virtual machines, whether it's Parrot, or the JVM, or now this MoarVM, has harmed the ability of Perl 6 to be implemented. We've seen one partial implementation after another, for years on end.
Unfortunately, we can't actually use any of these implementations for anything serious, like we can with Perl 5, Python, and Ruby. This makes Perl 6 unusable, which is quite a shame.
That's the second most disappointing part. The Parrot developers had a strong push between two and three years ago to redesign the internals to be a better fit for Rakudo along these lines (better runcore, easier JITting, improvements to the core object model, representation polymorphism), but the reaction from the Rakudo developers was a mixture of "No, thank you!", "Not yet.", and "If you have to, but don't change any APIs or behavior that Rakudo counts on."
It's a shame this couldn't have all been done at the same time. The available volunteer knowledge and interest might have saved two or three years of Perl 6 floundering around in the netherworld of big promises and very modest realities.
The most disappointing part is realizing that this sets back the potential delivery date of a usable Perl 6 by at least another couple of years. I guess that shouldn't be a surprise by now though.
Perl 6 has to target a VM, just as Perl 5, Python and Ruby do. The Rakudo Perl 6 team picked the Parrot VM and tried to focus on everything but the VM. That's clearly no longer tenable because Parrot has turned out to be a Norwegian Blue (at least at the moment).
It's disappointing that Perl 6 still isn't ready for prime time. Fortunately Larry Wall and friends are changing that. I'm confident MoarVM is yet another big leap in the right direction.
You may be trying to make a point with the account name, or may think you may not be able to get a fair conversation without the pseudo anonymity, but frankly, (I would hope) it's beneath you.
I would be happy to discuss my views on what useful and production mean. I'm pretty sure I know your current views on the topic (as a long time reader of your blog, my recollection is that they seem to have shifted over time).
Then again, maybe you aren't interested in having that discussion with me. If so, this seems a fairly inefficient way to communicate to the developers.