OK, here's the deal: if Perl 6 would have better designed, less bloated with superfluous features, and backward-compatible with Perl 5, I would not have started Rindolf. In fact, I consider Rindolf an anti-thesis to Perl 6. Thus, don't wonder if I criticize Perl 6 a lot, because it is the reason for starting Rindolf in the first place.
Perl 5 is a good language. Yet, it is facing a problem: competition. Perl faces competition from Python, PHP and to lesser extent Ruby, Tcl and other less dominant languages vowing for the so-called "scripting languages category". There's nothing wrong with this competition, and if someone is content programming in another language, there's no reason to tell him to use Perl instead. Still, it would not be good if Perl 5 lost ground due to the fact that it lacked features, which other languages possessed in a straightforward manner.
Now, Perl 6 is the wrong solution for the problem. As Perl 5 served as a handy scripting, glue and integration language in the past years, which saw an explosion of the Internet, of Linux, and of computing in general, there are now millions of lines of Perl 5 code written by hundreds of thousands of programmers. (perhaps even milliards LOCs written by millions of programmers - it's hard for me to approximate). These pieces of code do their job, work very nicely and are sometimes tweaked and modified as time goes by.
What is present in CPAN is just the tip of the iceberg. We can expect many modules in CPAN to have ad-hoc code that does the same job in many other places around the world. (I know I wrote a lot of code, that I later found out that there is a CPAN module that does exactly that). Furthermore, and equally importantly we can expect that there are now millions of Perl 5 programmers of varying skill and knowledge who constantly maintain such code or write new one for their own purposes.
Now, here comes a rhetorical question: exactly how are all of them going to convert to Perl 6, which is incredibly different and not compatible with Perl 5?
The answer is they are not going to. They are going to stick to Perl 5. Converters, you say? They are not going to work for several reasons. First of all, the programmers themselves are used to Perl 5. As we all know, everyone has his own Perl 5 style which is different than others. [1] Teaching them all Perl 6 is a recipe for failure. Secondly, a converter will probably convert your code to something crude and hard to manage. (if it will convert it correctly at all - eval "" anyone?). Lastly, people will be very reluctant to use the converter, if the code is already working in Perl 5. "It's working - don't touch". Perl 5 is probably going to be around for a while, so why bother converting a perfectly good code to the incompatible, new-fangled Perl 6?
And that's where Rindolf kicks in. Rindolf is a Perl dialect that is fully backwards compatible with Perl 5, yet offers some features not present there. Running "rindolf test.pl" will 99% of the time yield the same results as running "perl test.pl". A person can take a code written for perl5, use it as is with Rindolf, and gradually add Rindolf-specific features should he wish to.
Rindolf maintains backwards-compatibility because Perl 5 is good enough as a basis for a language. More importantly, it is a language that is in wide use today, and people depend on.
[1] And it is possible that a programmer writing the same script twice will not write the same script!