- From: Sailesh Panchang <sailesh.panchang@deque.com>
- Date: Thu, 21 Apr 2011 09:25:59 -0400
- To: "'James Craig'" <jcraig@apple.com>, "'Matthew King'" <mattking@us.ibm.com>, <jfoliot@stanford.edu>, "'Steven Faulkner'" <faulkner.steve@gmail.com>, <schwer@us.ibm.com>
- Cc: "'W3C WAI-XTECH'" <wai-xtech@w3.org>
Well sometimes the longdesc for a graph / chart may link to a page with a data table with the data points and a brief summary. Can that be done via an attribute like aria-description? What bothers me is that user agents (browsers and AT) do their own thing. Regardless of what HTML or ARIA specs say they choose to support / not support something. Certain ARIA features work in Firefox but not IE8 with screen readers, differences in behavior exist for HTML4 attributes like title, scope, etc. So I'll say get user agents to come up to a uniform level and then gaps can be addressed by new attributes / features / specs to be defined / developed. Longdesc is meant for non visual access to info that is redundant for sighted users. And is required only a small fraction of images and some might regard it as a clumsy mechanism but it is a workable mechanism that is already defined and supported by some screen readers. So is it worth the trouble of defining something new and then hoping user agents will implement it? Will it result in a significant benefit to users/ developers? The challenge is not using this attribute or that but actually coming up with good equivalent text for every unique situation that is written in a manner that helps users. Sailesh Panchang www.deque.com
Received on Thursday, 21 April 2011 13:25:36 UTC