We've tended to replace parts piecemeal; much of the underlying architecture has changed *dramatically* in the last 8-9 years but there's never really been a single cut-off point for "the whole thing" -- even when we swapped out cur/old for page/revision/text, or supplemented/started replacing the entire JavaScript infrastructure, large swaths of the code remained as they were, and there was often a lot of compat assistance.
As for potential future things that might earn a 2.0 label without a "full" rewrite, key things I'd look at that are likely feasible include:
* a MAJOR user-interface redo, moving towards an active front-end that communicates with the backend over API. A heavily HTML5/JavaScriptable frontend that can do offline work, use modern facilities for keeping a single "page" going while making using of history / URL updates to make page switching faster, more touch-orientation (eg like Brandon's "Athena" design mockup) so it scales up and down to large desktop screens, small desktop screens, small smartphone screens, and touch tablets.
Most importantly would be clean separation between frontend and backend: this would potentially affect every skin, every Special: page and a lot of custom code that works in the UI.
* major overhaul of storage/versioning/naming to support different datatypes like video or interactive programs really natively, branched versions, public and private drafts & sharing, etc
It may be that the former is easier to do that it feels, or that the latter might be less invasive than you'd think, but I suspect both will be lots of work at some point. :)
My suspicion is that we'll actually migrate our way slowly into the UI redo piece by piece, and we might end up splitting the other case into smaller pieces as well.
Whether a 2.0 apellation will be deserved at any given time? Honestly it just has to be "major" and user-visible; we don't have to get hung up on which piece at which time. :)
-- brion