> On Nov 20, 2013, at 10:42 AM, "Tab Atkins Jr." <[email protected]> wrote:
> 
>> On Wed, Nov 20, 2013 at 10:31 AM, Timothy Hatcher <[email protected]> wrote:
>> It does look good to me, though I can't evaluate it from an implementation 
>> perspective. I'll defer to others on that.
>> 
>> Requiring the <img> also also forces backwards comparability for clients 
>> that don't support <picture>. (As long as the author gives it a fallback src 
>> attribute.)
>> 
>> One part isn't clear to me yet: what <source> attribute is used for DPR 
>> multipliers? Allowing both multipliers and intrinsic widths would be ideal 
>> and flexible.
> 
> <source src> will contain a list of either url/density or url/size pairs.
> 
> ~TJ

Can it be srcset so <img srcset> is also usable for simple cases? (Yes respec 
srcset or anything not named src.)

— Timothy Hatcher

Reply via email to