Sociablecode

It’s no secret that many developers have been waiting in the wings for an OpenSocial RESTful API.

John Panzer has proposed an Atom/JSON standard. Give it a read and contribute.

So we’re a few months into OpenSocial here. I mean actual usable OpenSocial. Ok you’re right, we are just peeking around the corner and seeing useful APIs somewhere a dozen blocks away, but hey… it’s progress.

OpenSocial is open source. OpenSocial, as a community project, will yield not just a new social application platform, but an easily installable container that any website, CMS, isp etc can install in minutes and through which any of these will be able to provide what will one day be a thriving, exhaustive catalog of self-expressive, social applications.

Several months after the ballyhooed ‘release’ of the platform, which was essentially the equivalent of a debutante ball for an infant, most (almost all) of the ‘founding members’ of the movement are still basically waiting for someone else (read: google via orkut) to implement a reference container. What to think….

So does it in fact yield any advantage to be iterating simultaneously with Orkut and hi5 as they work through monotonous issues like adjustHeight() and sendMessage() and blah blah blah?

Yes.

Why? Because it will be hell on wheels to a) all at once decide how to implement all these necessary functions and b) track back through all the code decisions and namespace changes in order to help developers conform to the inevitable network-specific idiosyncracies as they arise.

Even using a common ‘container’ (i.e. Shindig), each network has:

- it’s own layout

- it’s own design

- it’s own messaging functions

- it’s own activity stream functions

- it’s own user data fields

…the list goes on, you get the point.

The point is, just because the API may change a bit between now and 1.0 or some other milestone where everyone will wake back up, Rip Van Winkle style and rush to install Shindig and remind everyone how on board they are with OpenSocial, each network still needs to, say, decide what they will allow to publish to the activity stream. Or, what a ‘message’ is when the sendMessage method is called, or any of a million other considerations,both programmatic and philosophical.

And let’s not forget the underlying necessity for good network-developer relationships.

The networks that have lots of usable, relevant, nice-looking applications will cash in the vig on that catalog readily. To this end, Google is actively saying ‘hey… don’t forget to make your Orkut apps look like, well, Orkut!!’ It’s a good point. If I want to write an app for my company for F8 or OS or whatever, it should look like part of the platform, plus our branding and value-adding features. The networks that are asleep at the wheel with OpenSocial will have a hell of a time getting everyone on board to make their apps relevant to their specific network.

That’s enough for now… more coding tomorrow.