[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

HTML5 and almost no namespaces

• From: Jesper Tverskov <jesper.tverskov@gmail.com>
• To: xml-dev@lists.xml.org
• Date: Thu, 2 Jun 2011 09:31:28 +0200

Considering that this XML List love to hate the details of "Namespaces
in XML", it is interesting to learn that HTML5 has more or less
abolished namespaces with a few exceptions.

They are still used implicitly (you don't need to declare them!) for
old applications like SVG and MathML, now part of the spec itself.
ARIA is also part of the spec in order to support it!

But for any future application that would like to make it into
webpages (probably many considering how the web is still evolving),
the rule is now clear: No more new namespaces. Any new application
must not be an application but must be new markup in HTML's XHTML
namespace. Anything new must be additional HTML markup. Period.

See Henri Sivonen's answer to my question at the HTML5 mailing list:
http://lists.whatwg.org/pipermail/help-whatwg.org/2011-June/thread.html

Namespace prefix and colon are in HTML5 part of local-name.

Forget about using XHTML5 served with "application/xhtml+xml" in order
to continue with the namespace mechanism. Browsers are never going to
implement new applications and markup if it is not also working in
HTML5 served with "text/html". At least that is my guess.

"Namespaces in XML" is far from perfect, but any mechanism allowing
for seamless co-existence of XML applications in the same document has
a price and is bound to be one of the most difficult aspects of XML
from a user point of view.

I feel that the HTML5 alternative to the traditional use of namespaces
is one big confusing mess. I find it hard to understand that W3C can
come up with a new HTML5 spec that to a great extent undermines core
concepts of XML with far-reaching ramifications.

At least HTML5 could come up with a better mechanism. Almost no
mechanism at all, or rather a 5% implementation of the traditional
namespace mechanism is a poor substitute.

Cheers
Jesper Tverskov
http://www.xmlplease.com

• Follow-Ups:
  • Re: HTML5 and almost no namespaces
    • From: Stephen D Green <stephengreenubl@gmail.com>
  • Re: HTML5 and almost no namespaces
    • From: David Carlisle <davidc@nag.co.uk>