Re: handling fallback content for still images

2007/7/9, Robert Burns:
> On Jul 9, 2007, at 2:35 AM, Thomas Broyer wrote:
> >
> > I don't understand what you're talking about.
> Earlier in this thread Mcaiej indicated that Opera and Safari (now
> anyway with an XML de-serialization) are adding an anonymous tbody
> element for CSS purposes.

Oops, sorry, I overlooked his message.

> This thread has evolved into basically
> discussing how to deal with subtle differences in the three forms of
> HTML5: DOM, XML and text/html. I think wit the CSS handled the way
> Opera and Safari are heading (with XML de-serialization) that takes
> care of many of the concerns I had. However, there still may be a
> need to address some subtleties in those particular elements.

However, XML+CSS does not necessarily mean XHTML+CSS, so a "pure"
XML+CSS renderer (with no special treatment for XHTML) won't generate
this "anonymous tbody", and the CSS will "break".
The above situation might be seen as an edge case but I believe XML
should remain XML and XHTML5 still is XML. In brief, I don't think
creating this "anonymous tbody" is a Good Thing: people have to
choose: either they use XHTML and they know it will be processed
according to XML rules (including applying CSS stylesheets with no
"anonymous elements" being inserted) or they use HTML. And if they
want to serve XHTML as text/html in some case and as
application/xhtml+xml in others, they just have to use "table > tr,
table > tbody > tr" (or just "table tr") as selectors in their CSS
stylesheets. It's not that hard after all (much less than having to
think that XHTML differs from XML with some "anonymous elements" being
inserted for styling purposes).

Thomas Broyer

Received on Monday, 9 July 2007 08:05:07 UTC