Re: XHTML Document Conformance (was Re: Unicode 3.2 Released)

Masayasu Ishikawa wrote:

> This issue was famous "validate against" discussion.  The discussion
> started in February 2000 from:
> 
>     http://lists.w3.org/Archives/Public/www-html/2000Feb/thread#51
> 
> and continued more than 6 months, including:
> 
>     http://lists.w3.org/Archives/Member/w3c-html-wg/2000JulSep/thread#423
>     http://lists.w3.org/Archives/Member/w3c-html-cg/2000JulSep/thread#116
>     http://lists.w3.org/Archives/Public/www-html-editor/2000JulSep/thread#200


I see. A huge can of worms.  But it's all about validation and what fell 
between the cracks is the simple issue that XHTML documents are supposed 
to first be XML documents, as the intro very strongly hints.  Nowhere 
does the document conformance section say that.

There should be an item 0 in the list of criteria in 3.1 that says: "An 
XHTML document must be a well-formed XML document as per [XML]". 
'well-formed' is redundant (if it's not well-formed it's not XML) but 
IMHO improves clarity.

Alternatively, the introductory sentence before the criteria could start 
"A Strictly Conforming XHTML Document is a well-formed XML document as 
per [XML] that requires only the facilities ..."

With such a statement, characters become properly defined (no need to 
refer to Unicode and 10646 anymore), as well as a whole lot of other 
things such as elements and attributes.  Right now there are definitions 
in section 2, but nothing linking them to a syntax.  You do not know 
that <p class="foo"> is a p element with a class attribute, it's undefined.


> and adopted the current workding in XHTML Basic and later, including
> the XHTML 1.0 Second Edition.  If we change the wording, it affects
> the whole XHTML Family.


I think it would be a Good Thing if the whole family inherited an XML 
well-formedness constraint that I believe was always on everybody's mind 
but just didn't get written down.

Regards,

-- 
François Yergeau

Received on Friday, 5 April 2002 10:12:34 UTC