Another Quiet Day

The Real Culprit is Google Authentication

Marc Canter said this here thing:

Specifics #5 = yes - clearly that’s called GData. And I feel so sorry for those 4,200 Apple developers - how do they think they’re gonna survive as slaves? But the shackles to keep me in place haven’t been invented yet. Or Dave Winer. So some of us can play this chess game just as good as the big boys.

--

This is in reply to Don Park's sharp as a blade observations on Google and Google's role in the current web application field. Don Park noted that the big companies will use APIs to rally the serfs to their flag.

However Marc Canter is wrong when he slams GData.

The real culprit is the Google Authentication API. They can use that API to close or open access to application data and control their integration.

The data API is unimportant from a strategy perspective so they've wisely used the most standard and open format for that API as is possible.

GData is nothing but a fairly sensible mix of OpenSearch and the Atom Publishing Protocol. Both are well-defined, designed for a fairly painless implementation. Their origin, although political to the very few A-listers who only see history as a piece of fiction for their egotistical grandstanding, is unimportant to the rest of the world.

Google Authentication, however, is a different matter. Authentication and identity is the integration point—the toll-booth on the road of communication.

Usually I'd find it odd that somebody who is a frequent proponent of OpenID and similar open specs for identity and authentication, then turns around and criticises Google for using Atom and the Atom Publishing Protocol (one at least on par with RSS, the other light-years ahead of XML-RPC in leveraging the basic scalability features of HTTP).

He doesn't mention a word on Google Authentication.

However, this is Marc Canter. Which explains the inconsistencies.

Baldur Bjarnason – 26/8/06