It got moved in the spec, because <video> fires resize events now, apparently. See http://html5.org/r/8354 Olli, any objections?
Huh, I'm surprised if firing resize on video doesn't cause regressions.
But I guess we can try this.
I don't see any spec bug about resize, not recent discussion on whatwg mailing list.
(and the regressions I'm thinking about are similar to what happened when normal event dispatch was tried for load event. But that didn't work out and load is special now.)
I'm not following what the concern is. Can you elaborate? The change was part of: http://lists.w3.org/Archives/Public/public-whatwg-archive/2013Dec/0096.html
In the case of load event the world expected that load event listener added to window was for the page load. And yes, there were tons of pages using capturing listeners. I wouldn't be surprised that if resize starts to fire for more event targets, we'll get similar issues as with load event.