John Milan has written an interesting piece about "social business applications" over on Read/Write web. He begins by defining "social business applications" as "software that coordinates group interaction that is important to running your business" -- a bit amorphous but workable definition. He then goes on to rightly identify "identity" as the key "feature" that is critical to social business applications. Further, he extends identity not only to humans, but to data as well.
All of this is fine and dandy, and John refers to identity mostly in its "attribute" forms (presence, etc.), but he really nails it when he correctly states that social business applications are all about "enabling the flow of data in such a way as to make its location immaterial." And now we're back in familiar territory.
Domain-based models of architecture have been predicated on using location as a proxy for identity. As we've begun to move to network-based models of architecture (the entire web 2.0/enterprise 2.0 movement), location is no longer an appropriate proxy. Accordingly, identity must become a core functionality of infrastructure. This distinction is key. Leaving identity at the application level (as John refers to it) actually causes a proliferation of the problems that not having identity causes. Abstracting identity from the application level has been the primary trend in all of identity management for the last ten years. As we move forward, abstracting identity out past the network and application layers will ensure that identity becomes an *infrastructural* component.
Without that move, social business applications can never realize their true potential.