If there are no successful, extant examples, then I don't think it's a good basis for anything. As I said, it's a Utopian ideal. That's fine, but it's not practical (and it's not trying to be).

I think there are going to be incredibly valuable lessons-learned from these federated models, and interesting things will happen there. However, all of that will be capitalized by a nimble, forward-thinking, idealistic-yet-centralized model that gets rid of all the ideological cruft and polishes a real, useful, and engaging product.