Again, I'd be pleased to be wrong about this, and hope the project succeeds, but it's not the approach I'd personally bet on.
None of the problems I'm talking about are insurmountable, but it does seem like it will be a continual source of inefficiency and friction for the lifetime of a project like this.