Re: http://www.w3.org/TR/PR-rdf-schema not valid SGML

At 02:27 PM 6/17/99 +0100, David J Woolley wrote:
>The DOCTYPE line is incompatible with the content, which
>means that I can't validate it with validator.w3.org.

http://www.w3.org/TR/PR-rdf-schema claims to conform to
the HTML4.0 "loose" DTD but in fact it doesn't.
http://validator.w3.org shows that the non-valid bits are all
due to the embedded RDF/XML markup.

>(My problem is that it breaks html2ps and html2ps requires
>valid HTML, so I really need to prove that it is 
>structurally valid before raising a bug report on html2ps.

The HTML4.0 DTD doesn't permit mixins with other XML markup.
As an experiment, we published all the RDF Working Drafts
with no DOCTYPE declaration and with mixed HTML and non-HTML
markup.  In fact, we made it well-formed XML.

When the time came to publish the Proposed Recommendations
we added a DOCTYPE declaration so that validator.w3.org could
point out all HTML errors but left the non-HTML markup, thus
introducing an inconsistency.

For the RDF Model and Syntax, when we published the final
Recommendation we moved the embedded RDF/XML data into
a separate resource.

http://www.w3.org/TR/PR-rdf-schema will be republished shortly.
We have the choice of making it well-formed XML once again
(and not usefully input to validator.w3.org), making it strictly
valid HTML (removing the embedded RDF/XML), or leaving it as
not-valid HTML-plus-RDF/XML.  Your input on the practical impact
of each of these options is welcome, especially in the context
of the work to reformulate HTML in XML; see "XHTML? 1.0: The
Extensible HyperText Markup Language"
http://www.w3.org/TR/1999/xhtml1-19990505

-Ralph Swick
 W3C/MIT

Received on Friday, 18 June 1999 15:25:15 UTC