Re: text/html for xml extensions of XHTML

On Sat, 16 Jun 2001, William F. Hammond wrote:
>
> I believe that Hixie, you, and I all agree that user-agents should
> not *for any reason* override content-types specified in HTTP
> transport. In fact, I previously pointed out that security problems
> can arise when this is done.

Ok, I'm glad we agree on that! :-)


> The definition of text/html is given in RFC 2854 and text/xml in
> RFC 3023.  Any UTF-8 encoded XHTML document may be served as text/xml
> under the text/xml spec.  XHTML, which is the XML version of HTML, is
> defined by http://www.w3.org/TR/xhtml1 .

Agreed.


> The text/html spec defers to the W3C specification for XHTML on the
> matter of when XHTML is permitted under text/html.

Correct.


> The exact extent of the overlap is not fully specified, and there
> is said to be continuing private discussion at W3C.

True, to some extent. However, the spirit of what _is_ specified is
that only XHTML documents that would be parsed equivalently by tag
soup user agents and XML user agents may be sent as text/html. Since,
by this definition, the behaviour of a browser parsing a valid XHTML
document sent as text/html as HTML tag soup cannot be distinguished
from the behaviour of a browser parsing the same valid XHTML document
as XML, there is no reason to complicate (and slow down) user agents
by making them sniff the content stream to decide which mode to use.

Also, by this definition, sending XHTML documents with mixed namespace
content (e.g. containing MathML) as text/html is incorrect, and UAs
therefore need not cope with his scenario.

-- 
Ian Hickson                                            )\     _. - ._.)   fL
Invited Expert, CSS Working Group                     /. `- '  (  `--'
The views expressed in this message are strictly      `- , ) -> ) \
personal and not those of Netscape or Mozilla. ________ (.' \) (.' -' ______

Received on Sunday, 17 June 2001 00:00:37 UTC