- From: Benjamin Hawkes-Lewis <bhawkeslewis@googlemail.com>
- Date: Wed, 12 Sep 2012 17:06:16 +0100
- To: Laura Carlson <laura.lee.carlson@gmail.com>
- Cc: "Edward O'Connor" <eoconnor@apple.com>, public-html@w3.org
On Wed, Sep 12, 2012 at 2:21 PM, Laura Carlson <laura.lee.carlson@gmail.com> wrote: > Another requirement would be to state in the spec that hidden content > MUST be discoverable to ALL users Seems unlikely we will get consensus on that requirement because: 1. We don't generally mandate how user agents expose semantics. 2. Such a mandate would risk breaching the Maginot Line between the HTML and ARIA specs PFWG is trying to police. 3. Some authors want to target content at particular user groups (for example, special instructions or controls for "screen reader users"), sometimes we treat this as a legitimate use case, and sometimes we treat features that act as a vague approximation of such selective provision as meeting that use case. 4. Sometimes we treat universal discoverability as met by a combination of mechanisms provided by client software and mechanisms coded by authors. Indeed, ARIA basically would not work without this combination. For example, JS developers build widgets with mouse and keyboard bindings then add ARIA annotations to communicate the semantics of those widgets to assistive technology. Consider the case where help for a control is hidden and referenced via @aria-describedby. The ARIA spec assumes that client software making use of ARIA semantics will provide a mechanism for discovering the help, but that users of mainstream client software will be dependent on mechanisms (e.g. focus event bindings) provided by the author. http://www.w3.org/WAI/PF/aria-practices/#kbd_tooltips -- Benjamin Hawkes-Lewis
Received on Wednesday, 12 September 2012 16:07:03 UTC