Gabe Wachob: Google's Open Social: The missing pieces? :

I see OpenSocial as being 3 pieces.

1) Gadget Writers can write gadgets for Orkut and they'll run on any social network that supports the openSocial specs. Hi5, Plaxo, Ning are already close. However, each SN Container will make it's own rules about what gadgets can run.

2) Social Networks that support the specs don't have to reinvent Facebook's API in order to deliver similar functionality. They can build a relatively rich profile page, relatively quickly and take advantage of a large ecosystem of gadgets by being a Container.

3) Social Networks and other websites can support the Data APIs. This should encourage the sort of ecosystem of Applications (web and desktop) and Mashups that has grown up around Twitter. But instead of being restricted to supporting one site, they should be able to work with any ste that implements the data APIs.

The problem today (7-Nov-07) is that:-
- The Orkut Sandbox is the only testing environment and it's flaky
- There are no examples of the data APIs anywhere.
- The specs for the data APIs are minimal and have some serious holes.
- The specs for the Container side haven't been released. And work on this is currently restricted to Google's NDA partners.

It seems that way too much of the early analysis has focussed on data portability between Social Networks. This is something that is only hinted at by the Data APIs. And at least for a while, the write-once, run-anywhere gadgets will be what is actually built. Except that they will actually be write-once, run-on-Orkut until there are other reliable Containers out there.


[ << Implications of Twitter providing OpenSocial Data APIs ] [ Ecademy now has support for the OpenSocial People Data API >> ]
[ 07-Nov-07 9:26am ] [ , ]