The mobile web is broken. Big time. This week at the Breaking Development Conference in Tampa, some of the smartest minds in web development gave presentation after presentation about missing APIs, unresponsive images and load times from hell.

Just as Marc Andreessen has said that “Software Will Eat the World,” on the web, applications are eating content. “Content is king,” yes, and we need to think about “content first,” but increasingly people are consuming content through applications.

So, as I have said, there is a showdown between open standards HTML5 for delivering applications on the mobile web and native apps for iOS, Android and Windows Phone. The flow of the future would seem to be going the way of HTML5, because as the device landscape gets more complicated and fractured, the cost of maintaing all of those different code bases becomes prohibitive, both in terms of time and money.

But, there’s a rub, in fact two. For the most part, HTML5 apps are slower than native apps, and in some cases, significantly so. In the world of mobile, optimizing for speed is the name of the game—see the success of Instagram and Tango, both native apps. And on top of that, mobile browsers do not provide device API support for many important features—especially the ability to access the camera.

As I speculated last week, Facebook’s Instagram buy is an attempt to maintain a competitive advantage in photo sharing through whatever means necessary while they wait for the HTML5 situation to improve. They may be waiting for a long time.

This is where the chess analogy comes in. If developing for the web is like three-dimensional chess to print’s two dimensions, then developing for mobile is like four-dimensional chess. But it turns out that the rivalry between Facebook, Apple and Google is turning this into HTML5-dimensional chess. Here’s what I mean:

Since Facebook launched its HTML5 application platform last year, they have placed major bets on the open web. They have developed a sophisticated HTML5 testing suite, called Ringmark, and released the code as open source. That they have twice as much traffic on their web app then on their native apps would seem to be a vindication of this approach. But the path to HTML5 apps is simple only if you ignore the important overlay of how these companies get paid.

Right now, Apple and Google get a 30% slice of any revenues, including in-app purchases, that run through their app stores. And consumers have shown a surprising willingness to pay a dollar or five for an app, but not the equivalent amount for access to a website that performs the same functions.

Facebook’s motivation for pursuing HTML5 is to play an end-around Apple and Google in the application market the way they have with their 30% take on the credits in the casual game market. Facebook may have the flow of history, and the ultimate best interests of their users on their side, but they are as commercially motivated as their opponents.

Along with releasing Ringmark, which defines mobile browser capabilities interms of a series of rings from zero to three with increasing levels of functionality, Facebook is courting developers to build “consensus” on “priorities” that the mobile browser makers, Apple, Google and now Microsoft, should build for. To that end, they have dispatched James Pearce, Head of Mobile Developer Relations, to present to developers, including those at Breaking Developement, “The real possibilities and opportunities that standardized device and network APIs can offer.” The goal is to convince them that, “Our hopes and dreams for a rich, contextual, social web will depend on them.”

Facebook, in this case, would seem to be playing white, but that doesn’t mean that they will definitively win any time soon.