On Dec 5, 2006, at 19:07, Henry S. Thompson wrote:
> So isn't the right thing to say about HTML5 that there _is_ a
> normative schema, but it's not for the surface syntax, it's for that
> underlying XHTML document.
If you count the requirements expressed in English as a "schema", yes.
> Seems to me there _must_ be such a
> grammar, or Hixie couldn't be figuring out what fixups to do: the
> fixups are designed to produce a tree of a certain form, and it would
> be useful for all of us to document that form using a schema. . .
The implied schema for the fix-ups is roughly HTML 4.01 Transitional,
because the fix-ups are based on browser behavior and browser
behavior has been influenced by HTML 4.01 Transitional. The English
"schema" for the higher-level HTML5 syntax is separate from the
parser fix-ups. There is a non-normative RELAX NG plus Schematron
schema in the works for the higher-level syntax.
--
Henri Sivonen
http://hsivonen.iki.fi/
