Re: handling fallback content for still images

Hi Smylers,

You're misunderstanding how this all relates to compatibility. The  
following quote is the best indication I see where I can intervene on  

On Jul 15, 2007, at 5:58 PM, Smylers wrote:

> Of course browsers can change their behaviour to meet our spec.  But
> note that exactly one of the following can be true:
> * We are recommending something that all mainstream browsers already
>   implement, so it is backwards compatible with them.
> * We are recommending that browsers change their behaviour to  
> implement
>   something new that wasn't in HTML4 and which browsers don't  
> currently
>   do.

This has go nothing to do with HTML4. This is about XML processing of  
HTML namespaced content. In XML (and HTML, but we're talking about  
XML) Visual UAs should replace the entire contents of  an <img>  
element with the sourced image. That means there is not  
incompatibility if an author places contents in that <img> element.  
In HTML4 that's not true since the invalidity error is typically  
treated as an ill-formedness error instead. But again, we're not  
talking about HTML4 or any text/html processing here. We're talking  
about xml processing.

My first point is to point out that we have an issue here with XML  
processing that we should address and not ignore. The most important  
point in addressing this is to try to maintain compatibility and do  
what's sensible. Its not necessarily sensible to allow authors to do  
this at all. However, it is sensible to guide UA conformance that do  
something with this content.

Again, this does not replace @alt at all in my view. So right now, it  
should be compared to @longdesc. Right now visual UAs are not  
handling @longdesc any better than the handle this issue.

I hope that makes things clearer.

Take care,

Received on Sunday, 15 July 2007 23:14:47 UTC