The OpenSocial blog has a good summary of Google’s recently announced Friend Connect in the context of OpenSocial:
… up until now, becoming a container – adding new social applications for your users – has meant having to provide your own source of personal and social information. By using securely authenticated APIs from existing social sites, Friend Connect means any website can host OpenSocial apps.
In the future, Friend Connect will call the RESTful API for containers that support OpenSocial v0.8, helping their users share their web-wide experiences with each other on their favorite social site.
Friend Connect uses three open standards to connect to other websites. It uses OpenID for identity and logging in, it uses OAuth to authorize access to friend and profile data on existing sites that host it, and it uses OpenSocial to embed the applications within your site.
Abstractioneer puts it like this:
In a nutshell, the OpenSocial RESTful API is a catalyst that enables participation in a much larger and more complex ecosystem.
What are the implications of this?
I think it punches wormholes to connect previously isolated sites into a network. That network is highly mediated by social identities and by contracts and permissions for social graph nodes, independent of the silos hosting such nodes. And it allows embedding of socially created artifacts into arbitrary hosts.
Such embedded social graph node representations can be said to be transcluded, so that their distributions across the network give it some sense of life.
Another implication is that Friend Connect makes identities more independent of sites – and this is exactly what OpenID supports as well. I’m me across multiple sites. And I can meet my friends across multiple sites. This is part of the ongoing Copernican Revolution of OpenID: user-centric identity.