Re: Validating XHTML with embedded RDF

Joseph Reagle <reagle@w3.org> wrote:

> Entities and well-formed internal subset parsing are a place where I easily 
> get confused, so I:
> 1. added my own entity declaration (<!ENTITY ent1 "Hello">) and used it 
> (&ent1) in an xhtml document; your schema considered the instance valid.
> 2. used an entity (&ent2) in an xhtml document that wasn't declared; your 
> schema *still* considered it valid.

It really depends on how exactly you wrote your document.  If you didn't
include any document type declaration and used an undefined entity,
it's not a validity error, but a well-formedness error.

  cf. http://www.w3.org/TR/REC-xml#wf-entdeclared

> Regardless, in trying to track what's still going on:
> 1. Just last month the Schema WG asked about using schema for 'XHTML + 
> MathML + SVG Profile' and you responded there's reluctance because of a 
> lack of entity support in Schema:
>   http://lists.w3.org/Archives/Public/www-html-editor/2003AprJun/0085.html

That's a secondary issue, and largely because of MathML, which has
much larger collection of character entities than XHTML.  The same
reasoning doesn't necessarily apply to RDF or XHTML 2.0.

> 2. However, you've also noted that XHTML2.0 will be using Schema?

I've been arguing to drop mandatory support for character entities
(and mandatory requirement to include a document type declaration
in a strictly conforming XHTML 2.0 document) from XHTML 2.0 for
a long time, partly because I believe supporting RDF is more
important than requiring support for a fixed set of character
entities.  It's been a long debate, and I have to say I'm not on
the majority side.

> If I understand that correctly, are these reconciled in an expectation that 
> something will come along as you work on XHTML2.0 that will provide you 
> with an equivalent of entities?

At the very least entitiy definitions should be separated from grammar
definitions.  Using a document type declaration doesn't necessarily
mean you have to use a specific document type definition.

Regards,
-- 
Masayasu Ishikawa / mimasa@w3.org
W3C - World Wide Web Consortium

Received on Thursday, 26 June 2003 03:44:44 UTC