Ruby 2.0 is a not a "start from scratch" project. It's an evolution of the language (which will be done mostly in the experimental 1.9 releases) with a new interpreter. Once 1.9 is stable it will turn into the 2.0 release. The actual effect on Ruby users will be noticeable, but minimal.
with one clarification... (last I checked) 1.9.1 will be the stable, it won't be incremented to 2.0. Matz decided against the even/odd = stable/dev versioning scheme.
Uh oh. I did not know that version 2.0 will be such a project. Any bets on whether it'll succeed at its goals by starting from scratch? The track record on that sort of thing in technology is notoriously lopsided in favour of failure.
1.8.x and 1.9 are just experimentation and intermediate steps to ruby 2.0 which is a "start from scratch" project