Re: handling fallback content for still images

On Jul 3, 2007, at 2:02 AM, Henri Sivonen wrote:

> On Jul 3, 2007, at 08:04, Robert Burns wrote:
>
>> On Jul 2, 2007, at 11:36 PM, Ian Hickson wrote:
>>> (In any case, for <video>, <audio>, and <canvas>, I don't think that
>>> separating these features from <object> makes things harder for  
>>> authors.
>>> If anything, I'd say it makes them simpler.)
>>
>> Would you say the same thing for <picture> ?
>
> FWIW, <picture> is significantly different because:
>  1) Still images already work *natively* in browsers and WYSIWYGish  
> editors using <img> since way back when.

video, audio, and flash all work "natively" in browsers and  
WYSIWYGish editors using <object> since way back when

>  2) Still images don't require an elaborate scripting API.

Not sure why that's relevant. They wouldn't require elaborate  
scripting APIs using <picture> either.

>  3) From an authoring perspective the marginal cost of switching  
> from <img> to <picture> is incredibly bad considering the marginal  
> utility. (To go from a string fallback to markup-enable fallback,  
> you break the compatibility with over a decade of software for the  
> non-fallback content. And normal authors care about non-fallback  
> more.)

I'm an ABD economist and I don't even know where to begin to respond  
to this (maybe once I finish my dissertation I'll understand what  
you're saying).

The vision here is that in the distant future no one will even touch  
<img> they will simply use <picture>. There won't b e any costs  
(marginal or otherwise) to using <picture> when compared to <img>.  
They will simply use <picture>. Hopefully they won't even know <img>  
was ever used unless they're a web historian. The benefits will be  
that the language will simply have one, easy to remember, element for  
still images that every author can use everywhere that supports easy,  
semantically rich fallback content.

<img> simply doesn't do that. @longdesc doesn't help because authors  
do not know how to use it. We can try to educate them in the short- 
term, but I don't think we should shy away from fixing the long-term  
problems simply because they'll take a long-time. It's not like if we  
wait a decade to try to fix these issues they will take less time.  
Rather if we fix them now, in a decade they might just well be fixed.

Take care,
Rob

Received on Tuesday, 3 July 2007 07:16:29 UTC