I met Will last month at an Austin PM meeting and had a rather lengthy discussion with him about RPerl. It is my impression the he has introduced those terms, and they are an attempt to differentiate between Perl operations that have direct counterparts in C++ vs. those that require additional massaging.
I agree that RPerl might be valuable for new projects that require as much speed as possible (think major web apps) but I suspect that it will be impractical in most cases to re-code existing any but the smallest existing projects. This seems to be especially true when one considers all of the popular CPAN modules that would need to be re-coded if one expects to see any significant speed boost in code that uses them. (That is not to say that many of the most popular modules would not benefit from such a make-over if RPerl lives up to its potential.)
It helps to remember that the primary goal is to drain the swamp even when you are hip-deep in alligators.