Lists Home |
Date Index |
On Aug 11, 2005, at 20:29, Robin Berjon wrote:
> Dave Pawson wrote:
> Mandatory DTD can be up for discussion given how it's been used in the
> past for quirks mode detection and the such, but for sure if there's
> any way to do without it, I sure won't mind.
I consider DTDs harmful in the browser context in particular, because
none of the well-known browsers that support XHTML 1.x actually process
the real DTDs, processing the DTD would make no sense from the
performance point of view, and pretending that DTD-based infoset
augmentation happens when the DTD is not really processed is even worse
than DTD-based infoset augmentation. And the whole point why the XML
spec allows DTDs not to be processed is that the writers of the spec
considered DTD processing incompatible with the way browsers work.
Of course, the doctype sniffing argument is and should be totally moot
when it comes to XHTML 2.0.
>> I'm with Uche on this one.
> As am I, except not entirely for the part where he talks about
> changing the namespace. Knowing when to change a namespace is a tough
> call and generally an unsolved problem, and calling it inconsistent is
> to overlook a few details.
From XHTML 1.x to XHTML 2.0 there is a lot of stuff that remains
relatively the same when considering the local element names without
the namespace. Changing the namespace is like running Rot13 on those
element names. Running Rot13 on element names like that is impractical
and uncool, because it makes code reuse in UAs more difficult and
prevents incremental adoption of new features in authoring.
I think taking the good ideas from XHTML 2.0 (to the extent they exist)
and backporting them to the XHTML 1.x namespace makes much more sense
in the context of the Web, IMO. That's one of the things that WHAT WG
is doing with the XHTML flavor of HTML5.