- From: Jason White <jasonw@ariel.its.unimelb.edu.au>
- Date: Sat, 30 Apr 2005 10:52:45 +1000
- To: "Gregg Vanderheiden" <gv@trace.wisc.edu>
- Cc: "'Loretta Guarino Reid'" <lguarino@adobe.com>, <w3c-wai-gl@w3.org>
Gregg Vanderheiden writes: > > Proposal 2: Replace "every element of the Web content" with "every user > interface component of the Web content", > > GV: GOOD > > then define "user interface component" as any part of the Web content that > can receive focus and accept input from the user. > > GV: NOT GOOD. User interfaces also provide information. You need to > either change this to 'user input interface component' or change definition > to: "any part of the Web content that conveys information to or accepts > input from the user". It's not as clean but we can't define user interface > as only input. Here's the problem: 1.3 already requires structure to be separated from presentation, which in a markup language means providing explicit structural elements. In general, this covers all of the elements used for output, that is, for presenting information; and by requiring label, role, state and value information we are now extending this to focusable user interface controls. Perhaps this is implicit in 1.3 already but it deserves special mention in the success criteria for the sake of clarity and completeness. If we require role information for elements that present information as well as those which accept input, isn't this tantamount to requiring it of all elements? If so, then there's a huge overlap with the requirement to separate structure from presentation. What would be the role information for, say, a paragraph element, and how would this differ from, or supplement, the element type (namely paragraph)? If we restrict all of this to user interfaces then we will have to define what a user interface is for purposes of this success criterion. Perhaps we could say that the success criterion applies only to elements/components of the delivery unit that either accept input from the user, or change dynamically in response to user input or external events. This is where role, state and value are meaningful, whereas they aren't in the case of a straightforward document or SVG image or other such static content. Thoughts?
Received on Saturday, 30 April 2005 00:53:40 UTC