Re: [whatwg] The problems with namespaces in text/html

Henri Sivonen <hsivonen@iki.fi> writes:

> On Nov 4, 2006, at 13:07, Elliotte Harold wrote:
>
>> Ian Hickson wrote:
>>> On Mon, 9 Oct 2006, dolphinling wrote:
>>>>> I'm not saying don't add MathML to HTML. I'm saying don't add
>>>>> namespace syntax to HTML.
>>
>> As I've said elsewhere, I find this viewpoint simply
>> incomprehensible. Namespaces are ugly, but they're not that ugly or
>> that problematic. They are also at the core of XML processing.
>> Throwing out namespaces makes it  nearly impossible to process HTML
>> with the very large and powerful set of XML tools.
>
> Note that what you quoted above was not about throwing away
> namespaces but about not introducing namespace *syntax* to the text/
> html serialization. In fact, HTML5 requires UAs to put HTML elements
> in the XHTML namespace.

Yes, I believe Ian Hixie's main point was not to use namespace
prefixes in html5.  If a content provider attempts a namespace
declaration using something like 'xmlns="http://www.w3.org/1998/Math/MathML"',
it should "work" in a user agent that handles the document under the
current xhtml+mathml regime (perhaps cued by an xml declaration [an
unknown PI for html5] or an xmlns attribute on the root 'html'
element) and also in a user agent that follows the new html5 way
(which presumably would ignore 'unknown' attributes).

This is the way to keep as much as possible of presently circulated
html-with-MathML content alive.  And much of it would then be servable
through http either the old way or the new way.  Also this way a user
agent like Opera which is very good at CSS might eventually be able to
handle such documents without knowing the MathML element vocabulary
once there is sufficient future development of CSS.

                                    -- Bill

Received on Saturday, 4 November 2006 22:57:12 UTC