- From: Jim Jewett <jimjjewett@gmail.com>
- Date: Mon, 24 Nov 2008 10:43:37 -0500
- To: "Henri Sivonen" <hsivonen@iki.fi>
- Cc: "HTML WG" <public-html@w3.org>
On Mon, Nov 24, 2008 at 2:29 AM, Henri Sivonen <hsivonen@iki.fi> wrote: > On Nov 23, 2008, at 06:49, Jim Jewett wrote: >>>> But <h1> and <div> won't need the processing >>>> spec. >>> So where would you put the requirements currently >>> in the "4.4.10.1 Creating an outline" section? >> In the processing spec. I don't need to know that to >> create a conforming document. > How do you know what an <h1> *means* in terms > of its level without running the outline creation > algorithm? (<h1> previously meant a top-level header. It created an implicit section break.) Now h1 means a header at the level of the current enclosing sectioning tag. It still creates an implicit subsection. The suggestion to not skip levels (replaced by the suggestion of always using h1) is good informative advice, but not needed to understand the meaning of the element. All the information about the relationship to other h* elements, and null sections, etc... isn't really needed to define the vocabulary. It is needed for the DOM. It is needed to generate identical outlines (a processing task) in some corner cases. But it isn't needed to define a properly used h1. [The extra information on h1 in particular is indeed something that could *probably* just be handled by a subset view. I don't believe that is the best solution, but agree that if it proves inadequate, h1 isn't likely to be where the problem arises.] -jJ
Received on Monday, 24 November 2008 15:44:13 UTC