Re: Feature queries

On Dec 10, 2009, at 8:53 AM, Anne van Kesteren wrote:

> On Thu, 10 Dec 2009 17:28:47 +0100, Brad Kemper <brad.kemper@gmail.com> wrote:
>> I like it. Of course, I am one to say publicly that browser detection is useful (which I know is contentious). Like this:
>> 
>> 	input#userID  { background-image: url(images/label-UserID.png); }
>> 	@supports ( -webkit-box-reflect:below ) {
>> 	    	/* webkit UAs only: */
>> 	    	/* (nothing to do with box-reflect) */
>> 		input#userID  { background-image: none; }
>> 	 }
>> 	input:focus {background-image: none; }
> 
> I think this demonstrates why this is a bad idea. It will be much harder for WebKit to phase out support for the prefixed version of box-reflect after hacks like this get deployed.

You could say that about any use of -webkit-box-reflect though. Apple is not likely to phase it out soon if it is in widespread use. And of course, I only picked -webkit-box-reflect semi-randomly. I could just as easily have used '-webkit-border-radius', which may never be phased out.

There is also the point I made that you didn't quote, which is that such a hack would be needed much less if we could also determine support for attribute-value pairs.

I feel it is NOT a bad idea. It is an invaluable idea that has been a long time coming. i suspect it is actually _vital_ in order to get significant authoring with new layout properties like template layout.