That's why this is really a distributed systems problem. In order for an app-developer to build such an app, the deeper problems around deployment, sync, and identity/connectivity need to be resolved. Developers just need something to target that provides the appropriate components. That's what the toolstack I previously linked is trying to solve. Without that, we're stuck with requiring someone else to provide the central system/aggregator. Clearly this is a non-trivial problem but if it can be cracked then it'll be as impactful as the opening up of the Web was.
Thinking about this, I wonder if the economics would work given the ad-supporter/loss-leading aspects of some of these services. Still, interesting.
I supposed it could be, or a UI problem....it all depends on the architecture of the code and how you choose to define the problem. Regardless, it's all certainly feasible and possibly will be done sooner or later, it's just a matter of when. Are you familiar with the Diaspora project from a few years back?