It’s possible to create new HTML elements, and apply presentation and action to it.
Jon Resig mentions it here, suggesting it would be a good way to implement HTML5 in an earlier browser.
He’s right. It would be good. But I can see applications of it for microformats and semantic markup too.
If you attach behaviors (using Dean Edwards’ Mozilla extension) to the element, you can add script actions as well as style.
So you can essentially rewrite the browser using JS.
The accessibility crowd will call me out, since you’d need JS. But if your markup is that much more readable, it shouldn’t bamboozle the screen-readers etc. Hell you can even support that 1% we always detect who’re browsing using telnet in a green screen uni lab.
Eg, this site could go like this:
I know I could just do that with XSL anyway, but only geeks like XSL. 😛
Prize to the first person making IE7 pass Acid2 with just JS.