Re: Let every element have a src attribute

On 4/3/07, Dao Gottwald <dao@design-noir.de> wrote:
> Jeff Schiller wrote:
> > But to me, the biggest argument against the src attribute everywhere
> > is this:  Let's say authors start heavily using it - now suddenly
> > older user agents, which might have been able to display the image in
> > the first place, will be ignoring the src attribute and only
> > displaying the fallback text.  Why?  Suddenly, pages that could have
> > still rendered in older user agents will stop displaying their images.
>
> That's the minimum price that you have to pay when evolving a spec.
> The web site won't break in old user agents, which is a real advantage
> over most other changes that will make HTML5.

I agree that the website won't break, but my point was that a website
that is perfectly capable of displaying the image, will not because of
the new markup that authors will want to use (or conversely, src as an
attribute will not be used for fear of messing up older browser
renderings).

>
> > The HTML:object already exists for this exact fallback mechanism, we
> > should be clarifying any unspecified behavior of the HTML:object
> > element, not introducing new attributes/elements that do the same
> > thing but "break" older user agents.
>
> Old user agent won't magically fix their object implementation. And
> because they don't necessarily support the fallback mechanism, this
> really breaks your site.

Ok, it's possible I need a lesson on which browser agents don't
support the object fallback mechanism, but in a simple test like:

<html>
<body>
<p>This is paragraph 1.</p>
<object src="santa.claus" type="burger-meister/meister+burger">
  <p>This is paragraph 2, it contains <a href="http://www.w3.org/">a
link</a>.</p>
  <p>This is paragraph 3.</p>
</object>
<p>This is paragraph 4.</p>
</body>
</html>

The markup is displayed as expected in:

- Internet Explorer 6
- Firefox 2
- Opera 9
- Konqueror 3.5.5

Can someone check Safari?

I realize that this is the lion's share and that the edge cases here
might make the difference, especially for accessibility, but if most
of the existing HTML4 browsers support object's fallback mechanism now
- I say we should use that mechanism in HTML5 as well.

Jeff

Received on Tuesday, 3 April 2007 13:34:13 UTC