Re: Normative status of author-only view of the HTML5 specification

The authoring document should not define browser behavior -- that
was the whole point behind it.

Jonas Sicking writes:
 > On Fri, Jun 10, 2011 at 1:50 AM, James Graham <jgraham@opera.com> wrote:
 > > On 06/10/2011 02:35 AM, T.V Raman wrote:
 > >>
 > >> Paul, you are correct. If the two documents disagree after the
 > >> one is a projection of the other, then an cinsistency points to a
 > >> much deper problem.
 > >
 > > It seems much more likely to point to a minor error in the redaction process
 > > e.g. some content accidentally being excluded from the author-only view, or
 > > a bug in the scripts. In such cases one would always expect the main spec to
 > > be right and never the author-only view. Hence the main spec should be
 > > normative, the author only view should not, but should include text like
 > > Henri suggested indicating that it is a reduction of the normative spec and
 > > discrepancies should be reported as bugs.
 > 
 > Another thing that absolutely will happen, and which should also be
 > clarified, is what happens when one of the two documents leaves
 > something undefined, but the other document does define behavior. It
 > will most likely be very common for the authors document to leave out
 > details, or simply state that something is not allowed, but have the
 > HTML5 spec define very strict required behavior which affects both
 > authors and implementations.
 > 
 > So we should state that if anything is left undefined in the authors
 > document, but the HTML5 spec defines normative requirements, then the
 > HTML5 specs normative requirements are the ones that apply.
 > 
 > / Jonas

--

Received on Saturday, 11 June 2011 16:25:29 UTC