This doesn't seem to be linked to any upcoming or finalised W3C/WHATWG standard. It certainly looks like a nice idea, but isn't this arbitrary introduction of non-standard features exactly what we hated IE3-6 for? Prefixing it with "moz-" doesn't really make it less propriatory (we can easily have imagined all the IE* 'extensions' having an ie- prefix). If IE9 came out with a load of new CSS selectors and syntax that nobody had heard of, there'd be uproar in large sections of the technorati.
This kind of experimentation is certainly useful. However, with the vendor- prefixes leaking into real world sites, we seem to need a real playground for experimental new CSS/HTML features that ensures they don't get used in production. Alas, I'm not sure what the best route to this would be.
The vendor-prefixed properties being used on production sites are the fault of those who create the sites, not those who create the browsers. The purpose of the vendor-prefixes is well known: to put new capabilities into the hands of developers and designers to allow them to test different implementations without having to use non-production versions of browsers. You are welcome to use these properties however you see fit, but you don't get to complain when they go away or change implementation because they are, by definition, not standardized and in-flux. So if your site is relying on them, and you don't want to spend the time to maintain them when they change, it's your own fault.
More than any added fancy gradient, shadow, etc. property, I look forward to progress on the CSS syntax. Supposedly it is so simple as to allow non-programmers to use it, but I honestly have never met a non-programmer who found CSS easy to use. It would be nice if we can keep the old syntax around, but have shortcuts like these added for those who wanted to use them.
OT but my favorite item on his blog is this one http://dbaron.org/log/20100309-faster-timeouts which helped me figure out an animation smoothness problem was with the browser renderer itself and not settimeout specifically
This kind of experimentation is certainly useful. However, with the vendor- prefixes leaking into real world sites, we seem to need a real playground for experimental new CSS/HTML features that ensures they don't get used in production. Alas, I'm not sure what the best route to this would be.