[whatwg] Outline Depth Does Not Correspond to Sectioning Depth?

Mon Mar 26 17:08:21 PDT 2012

On Mon, Mar 26, 2012 at 5:02 PM, Hugh Guiney <hugh.guiney at gmail.com> wrote:
> I am confused by the current definition of outline depth (
> http://www.whatwg.org/specs/web-apps/current-work/multipage/sections.html#outline-depth),
> which, if I understand it correctly, states that the depth resets with each
> sub-outline.
>
> So, in the following:
>
> <body>
>  <h1>Document Title</h1>
>  <section>
>    <h2>Section Title</h2>
>    <section>
>      <h3>Subsection Title</h3>
>    </section>
>  </section>
> </body>
>
> …each heading would have an outline depth of 1, yet for:
>
> <body>
>  <h1>Document Title</h1>
>  <h2>Section Title</h2>
>  <h3>Subsection Title</h3>
> </body>
>
> …which is, according to the spec, semantically identical to the previous
> example, each heading would instead have an outline depth of 1, 2, and 3
> respectively.
>
> At least, that is how this implementation (
> https://github.com/hoyois/html5outliner) behaves; I raised this issue with
> the implementor and he seems to think it is the correct behavior—if so,
> why? Wouldn't it make more sense to have the depths be 1, 2, and 3, whether
> explicit sections are used or not?

I don't understand.  The relevant spec text is:

"The outline depth of a heading content element associated with a
section section is the number of sections that are ancestors of
section in the outline that section finds itself in when the outlines
of its Document's elements are created, plus 1. The outline depth of a
heading content element not associated with a section is 1."

This seems pretty clear - the outline depths of the headings in the
first snippet are, respectively, 1, 2, and 3, since that's how many
sections (plus 1) are the ancestors of each.  The same applies to the
second snippet, since the implied sections generated by the headings
produce the same virtual structure as the first snippet.

What text are you looking at that implies the outline depths of the
headings in the first snippet are all 1?

~TJ