W3C home > Mailing lists > Public > public-html@w3.org > November 2010

hgroup and ARIA outline in ISSUE-129 [Was: SSUE-129: replace or modify the ARIA section of the HTML5 spec - chnage proposal and replacement spec text]

From: Benjamin Hawkes-Lewis <bhawkeslewis@googlemail.com>
Date: Mon, 22 Nov 2010 07:59:15 +0000
Message-ID: <AANLkTinF-FNRB-HQosg_8KXodNbiHNFCiLCB=Zr0qSYn@mail.gmail.com>
To: Steve Faulkner <faulkner.steve@gmail.com>
Cc: HTMLWG WG <public-html@w3.org>, Maciej Stachowiak <mjs@apple.com>, Paul Cotton <Paul.Cotton@microsoft.com>, Sam Ruby <rubys@intertwingly.net>, Richard Schwerdtfeger <schwer@us.ibm.com>, Cynthia Shelly <cyns@microsoft.com>, "Michael(tm) Smith" <mike@w3.org>, Janina Sajka <janina@rednote.net>
On Sun, Nov 21, 2010 at 11:28 PM, Steve Faulkner
<faulkner.steve@gmail.com> wrote:
> change proposal: ARIA in HTML5: replace current text of section 3.2.6
> - http://www.w3.org/html/wg/wiki/ChangeProposals/ARIAinHTML5
> associated specification text:
> http://www.paciellogroup.com/blog/misc/HTML5/aria-html5-proposal.html

Under "basis for defining h1 to h6 element that does have an hgroup
ancestor", the Change Proposal states:

"It is not logical or practical that the semantic information about
both the heading elements contained within the hgroup and the hgroup
itself is conveyed to assistive technology."

Why?

The Change Proposal continues:

"In order to provide a heading structure to AT that is more sensible,
it is proposed that headings inside a hgroup have a default role of
presentation. This will result in only the one hgroup heading (instead
of 3) being reported to AT."

This introduces a significant difference between the outline algorithm
specified by HTML5 and any outline generated using ARIA.

I'm not a colossal fan of how HTML5 repurposes the old hX elements,
but I think it should at least be consistent between HTML semantics
and the accessibility tree.

In HTML5 only the first highest ranked heading contributes to the
"hgroup" text; in this proposal all headings do.

So given:

<hgroup>
  <h2>Foo</h2>
  <h1>Bar</h1>
  <h2>Baz</h2>
  <h1>Quux</h1>
</hgroup>

in HTML5 the heading is "Bar" but in the proposal in the accessibility
tree the heading is "Foo Bar Baz Quux" which subverts author intent.

Applying "role='presentation'" would also block UAs from exposing the
actual semantics of "subtitle" that lesser-ranked headings possess,
were any accessibility API to include that semantic, since "For any
element with a role of presentation and which is not focusable, the
user agent MUST NOT expose the implicit native semantics of the
element (the role and its states and properties) to accessibility
APIs".

http://www.w3.org/WAI/PF/aria/roles#presentation

I think this would be a more accurate reflection of intended semantics:

<hgroup role="heading" aria-level="whatever" aria-labelledby="heading-text">
  <h2>Foo</h2>
  <h1 id="heading-text">Bar</h1>
  <h2>Baz</h2>
  <h1>Quux</h1>
</hgroup>

Spec text could be along the following lines:

"hgroup element: heading role, with the aria-level set to the
element's outline depth and the aria-label, and with the first
highest-ranked heading element providing its accessible name."

"hX element that does have a hgroup ancestor: No role."

This would also "result in only the one hgroup heading … being
reported to AT" when listing headings, and would allow UAs to
represent subtitles to supporting accessibility APIs.

--
Benjamin Hawkes-Lewis
Received on Monday, 22 November 2010 07:59:54 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:21 UTC