On Tue, 21 Feb 2012, Jonas Sicking wrote: > On Tue, Feb 14, 2012 at 10:05 PM, Ian Hickson <[email protected]> wrote: > > In short, I agree that if the implementation cost was high here that we > > could compromise on this design and go with something less clean or with > > less graceful fallback, because it is true that many authors will not use > > this fallback feature. However, given that the cost is low, I don't see > > why we would remove the fallback feature. It's relatively simple, > > inobtrusive, and works. > > The fact that a feature is easy to implement should carry very little > value.
It does. But it's not zero value. > We shouldn't add features unless they are going to be used. I'm > personally very unconvinced that this feature would be used enough to > warrant adding it to the spec and implementing it in browsers. I reach the opposite conclusion, but I agree that it is a judgement call for which we have as yet insufficient information to be confident one way or the other. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
