Re: Relation between markup and transport

Ian Hickson <ian@hixie.ch> Mon, 2 Jul 2001 16:12:17 -0700, writes:

> > 1. More reliable CSS styling.
> 
> Can you give me an example of a page written in XHTML that is
> rendered more reliably than a page written in HTML 4?

Surely, we're not discussing the behavior of a particular rendering
agent here.  But for example, an incorrect HTML document cleaned up by
Dave Raggett's "tidy" might indeed show up better.

> > 2. Namespace extensions.
> ...
> section 5.1 of XHTML states that only documents that, by virtue of

Section 5.1 explicitly allows as text/html a doc prepared consistent
with appendix C without forbidding any other XHTML.  The
recommendation provides a specfication for XHTML 1.0 without fully
dealing with the issue of the relationship between the markup it
describes and transport content type.

> > Remember that W3C's Amaya handles MathML under either transport
> > content type.
> 
> And does so by using a heuristic that makes it unable to render valid
> HTML documents:
> 
>    http://damowmow.com/mozilla/html-not-xml.html

No.  Amaya, as an XHTML-aware user agent, complains correctly about
the presence of an incorrect XML declaration.  The example fails also
as HTML 4.01 in regard to the provision of section 7.1 that a document
must begin with a doctype declaration.  It is only valid as an
instance of the SGML application canonically associated, under the
HTML 4.01 spec, with HTML 4.01.  It is a subtle point, but it is also
an important point.

> > A. Oxford-TEI-Pizza-Chef-custom-brew-with-math under XML. Serve as
> >    "text/xml". Browser provides tree portrayal if no external
> 
> ...or if the application can render the document natively, for example

Sure, whatever else the mass-market user agent might provide is OK;
the tree representation is a minimal expectation.  The point is that
such agents should not be allowed to "own" text/xml.  Such agents
should be required to respect a user's webcap or mailcap entry for
"text/xml" as for any other content type except possibly "text/html",
which by historical precedent is an exception that plays the role of
the web's default content type.

                                    -- Bill

Received on Thursday, 5 July 2001 15:52:54 UTC