Scrivener uses XML though. I’ve even had to dive into it in order to rescue a project eight years ago. Could the text of the binder objects be stored as XML or even just raw JSON? I’m sure it could be, but It might not be the cleanest conversion in the back of the house.
Sometimes legacy code is left alone for good reasons, especially if it’s still working at something approaching usefulness.
I don’t know enough of why RTF was the magic bullet when Scrivener came along, I just remember that it was the magic format for most digital submissions in the early aughties. I seem to remember that it had something to do with the way RTF stores instructions being dependably reproducible no matter what machine it’s on or which machine it was written with (which at the time Word was not always spot on between Windows and Mac versions).
While it might be interesting to hear the official word on this, I think I’d much rather the effort is spent on the new app, as I’m not in the beta and would love to finally get to play with it.