Jon Howell dreams of a new Internet. In this new Internet, cross-browser compatibility checking is a distant memory and new features can be unilaterally be added to browsers without having to convince the world to upgrade first. The idea which makes this Internet possible is so crazy, it just might work.
What if a web request didn’t just download a web page, but the browser too?
The answer is very little: all you need is a native execution environment, a minimal interface for persistent state, an interface for external network communication and an interface for drawing pixels on the screen (ala VNC). That’s it: everything else can be implemented as untrusted native code provided by the website. This is an interface that is small enough that we would have a hope of making sure that it is bug free.
What you gain from this radical departure from the original Internet is fine-grained control over all aspects of the application stack. Websites can write the equivalents of native apps (ala an App Store), but without the need to press the install button. Because you control the stack, you no longer need to work around browser bugs or missing features; just pick an engine that suits your needs. If you need push notifications, no need to hack it up with a poll loop, just implement it properly. Web standards continue to exist, but no longer represent a contract between website developers and users (who couldn’t care less about under the hood); they are simply a contract between developers and other developers of web crawlers, etc.
Jon Howell and his team have implemented a prototype of this system, and you can read more about the (many) technical difficulties faced with implementing a system like this. (Do I have to download the browser every time? How do I implement a Facebook Like button? What about browser history? Isn’t Google Native Client this already? Won’t this be slow?)