I wouldn't call this approach radically new, this was already done with GTK3: http://blogs.gnome.org/alexl/2010/11/23/gtk3-vs-html5/ http://blogs.gnome.org/ale

I wouldn't call this approach radically new, this was already done with GTK3: http://blogs.gnome.org/alexl/2010/11/23/gtk3-vs-html5/ http://blogs.gnome.org/alexl/2010/11/23/gtk3-vs-html5/ Also we already have a lot of browser UI toolkits that behave the same in modern browsers. I don't really feel this could bring anything new to the table. A lot of the time in fact we don't want our UI to look the same everywhere, for example I would love my site to scale down and degrade gracefully on mobile or in other words: progressive enhancement. Performance and legacy browsers still plague the canvas element, so I won't rely on that as the foundation of a UI technology. And last but not least: this UI framework has to duplicate a lot of behavior that the browser already does, which I feel is both a waste of time and a waste of resources. I don't want to look too negative about this, but I'm really skeptical that a UI framework like this would be better than what we already have.