Went to a talk about the web-kit animation stuff. It's groovy, but
very iPhone-centric, and not a lot of use outside of that platform.
Well, <canvas> was Mac-widgets-centric, and had not a lot of use
outside of the Mac OSX platform, but now it's built into FFs, Chromes
and Operas too.
The web can't and won't move forward if it sits waiting for the w3c
(OMG) to come up with the next great thing(s), that's why we need
experiments like this (and most of HTML5's, and Mozilla's, and
Google's), that might eventually be ~ widely recognized as a "good
thing", and implemented in most browsers. At that point in time, they
become a standard (at least a de-facto one).
You also have to be careful to keep your animations limited to those
properties that Apple decided to hardware-accelerate for the iPhone.
There are things lacking any importance, that, instead, when
developing for the iPhone become important. For example, a page with a
setInterval() when left open on the iPhone could easily drain the
battery in an hour, while the same page without the timer could be
left open ~ forever. You would not care about a timer's power
consumption in a desktop browser, as the power supply is not a
concern. The point being that each and every additional cpu cycle
spent in a job costs battery life, -which is not a good thing on a
mobile phone- and things like e.g. fading out an element via CSS
via .style.opacity=x; with a timer in a loop in JavaScript code are
more expensive in terms of power than a browser-native hardware-aided
alternative. On top of extended battery life, you get a smoother fade-
out effect too.