Dave Morin & Ari Steinberg from Facebook gave a preso a the Web 2.o expo that I have been thinking about all week. The question they addressed was "What is a great API and how do we build one for Facebook?"

Of course, the real question behind that question is "How do we grow tools for developers so we can build Facebook into an ecosystem that will beat MySpace?

Or, in other words, do do we take a destination that has (according to Morin) 1.5 billion PV a day, and 500,000 new profiles created a week, and fast-forward to a number of compelling applications that can be developed for inside the FB community?

We've already seen the FB team introduce expanded photos, notes, and invites; the goal now is to permit developers to build additional services for FBers to use inside the tent.

Morin's preso becomes especially interesting to me when Peter Cashmore talks about Facebook's interest in developing their own classified ads and when Mike Arrington rips the covers off MySpace's (still not live) new news service (like, why, ever leave, huh?)

It's even more interesting when you look back at Web 1.0 companies and the network effort of traffic; what did Netscape and MSN seek other than to provide a host of integrated services that would incent users to *never leave* and to circulate around the network?

For social networks with strong communities and page views, the opportunity is the same: breed'em, keep'em, multiply'em and if you can stick some ads on there--and sell some premium services--you've got a win.

Or, in other words, the paradigm for success today goes two ways: create APIs that will distribute your data out--but if you seek to intensify the network effect--also make sure to create APIs that will incent developers to help you keep your users in.

More to come this theme...I want to explore to what extent there is a strong correlation between early or sustained success (think sixapart) and partner integrations via deal and/or API (think Mozilla).