I released RTF::Tokenizer v1.0 last night. All hail me. I used FIGLET for the README file again...
It's a massive improvement. I stole the best parts from RTF::Parser and the old RTF::Tokenizer, and got a huge speed improvement. There are over three thousand tests too.
The plan is now to build RTF::Reader using it, and also rewrite RTF::Parser to use it, which is easier said than done, but, done it shall be.