This sounds like something we already have. I don't want this to sound confrontational, since I know the F-word is viewed as kind of "dirty" around here, but how is what you're proposing substantially different from Flash?
Then, this is based on doing websites in the web standards spirit of using plain-text files, where multiple components are linked together, and having the semantic content (XML) separated from the presentation (HSS).
I mean, we like the concept of HTML+CSS, but we would like to see many of its flaws and shortcomings corrected.