What is an Internet operating system?

I trace the phrase Internet operating system back to a 2002 essay in which Tim O’Reilly imagined that the Internet OS would arise from, and become the governing framework for, a soup of ingredients:

All of these things [including web services, p2p filesharing, blogs] come together into what I’m calling “the emergent Internet operating system.”

…
In the third stage, the hodgepodge of individual services will be integrated into a true operating system layer, in which a single vendor (or a few competing vendors) will provide a comprehensive set of APIs that turns the Internet into a huge collection of program-callable components…

Of course the web had always been a collection of components, as I had pointed out in 1996, but the implicitly-available services woven into the web’s fabric were hard to use back then, and in many ways still are. One key enabler for the Internet OS, therefore, would be a framework for defining and deploying services. Another would be universal data-exchange mechanisms that would supply the grease to overcome data friction. Still another would be standard ways for services to communicate through intermediaries that support authentication, authorization, and group membership.

The Internet itself, meanwhile, had always natively supported peer-to-peer networking, a capability that was eclipsed when its success spawned a layer of NAT (network-address-translation) firewalls protecting hordes of semi-connected private networks. As a result, enabling the new Internet OS would also require some means of restoring that original P2P connectivity.

What blogging brought to the table, in addition to the liberating power of personal publishing, was a new take on the venerable publish/subscribe pattern, expressed now in terms of the familiar metaphor of news syndication. In any version of the new Internet OS, syndication-oriented architecture would have to play a crucial role.

Fast-forwarding to 2008, I’ve been reading definitions of the Internet OS like this one from Doc Searls:

[Google’s] Chrome also runs apps. In that respect, it’s more than the UI-inside-a-window that all browsers have become. It’s essentially an operating system.

An Internet OS is, to be sure, a platform for running applications, though that’s a slippery term given application styles ranging from native-to-the-underlying-OS to dynamic-HTML-plus-JavaScript to rich-a-la-Flash-and-Silverlight. But when you expand the notion of an application beyond UI-inside-a-window, a number of supporting themes come into view: universal data exchange, peer-to-peer connectivity, group formation, publish/subscribe messaging, syndication-oriented architecture. One of the places where these themes come together is Live Mesh, as Mike Zintel explained eloquently in Live Mesh as a Platform.

As I mentioned in my interview with Ray Ozzie about Live Mesh, the Internet OS meme morphed into the Web 2.0 meme which embodied an application style based on dynamic-HTML-plus-JavaScript and a cultural preference for open participation. But it was inevitable that the original notion would come round again. As it does, it behooves us to ask: “What is an Internet operating system?” I think we’ll find that it includes what we mean by Web 2.0 but also expands that meaning in ways we’ll want to discuss and define.