Re: Discussing ARIA in HTML5 integration

aloha, dan!

thanks for the quick reply...

i'm not sure the technology is extant for the type of test cases you're 
asking for, as they presupposes an HTML5-capable renderer (or simulator 
of such using a snapshot of the working draft) and a fully ARIA-aware 
assistive technology (or simulator of such a beast)...

granted, some simple simulated proof-of-concept tests could (and should) 
be run, and several keyboard support features should be tested without 
assistive technology, for an important, and often over-looked 
constituency of users who benefit from "native accessibility features" 
in both markup languages and the user agents which render them, and a 
co-operative underlying operating system in order to have any meaningful 
interaction with a computer, but who do not require an "assistive 
technology" [reference 1]

but, since ARIA 1.0, as currently deployed (and described in the ARIA 
Roadmap [reference 2], provides a means of enabling interaction with 
markup by conveying relationships, states, properties, roles, and events 
to accessibility APIs, the problem of simple manual testing of 
ARIA-enabled HTML5 pages is not quite such a simple task, although the 
concerns of this population is one of the focuses of the User Agent 
Accessibility Guidelines, version 2.0 [reference 3] and a major motivator
for the inclusion of as much native accessibility support in HTML5 as 
possible...

as for interaction with on-screen keyboards, speech input, tactile 
displays, and speech output, for example, those who actually use the 
technologies and know how to evaluate their efficacy would need 
continually updated access to the very latest alpha builds of commercial 
assistive technologies, due to the predominance of third party assistive 
technologies developed by for-profit companies, despite such laudable 
initiatives as NVDA [reference 4] Orca [reference 5] and OATSoft 
[reference 6]

charles, does this hold true for FireVox [reference 7], since FireVox 
runs in FF2 and not FF3?  chaals, does this hold true for Opera with 
the XHTML+Voice plug-in? [reference 8]

yes, there is something to be gleaned from theoretical and/or automated 
tests, but there is a limit as to their efficacy, for -- as we all too 
well know -- one can construct a valid page that is unusable by 
anyone...  

i think our (collaborative, cross-working group) efforts will bear more 
(and more immediate fruit) if (1) PF concentrates on getting ARIA 1.0 
to recommendation status, (2) members of the HTML5 WG with the expertise
and inclination (such as HenriS and SimonP, whose efforts and persistence 
i laud) assist the PF WG to gather as much hard data as to an integration 
strategy which will work in HTML5 without upsetting anyone else's apple 
carts, and (3) pertinent WAI working groups and HTML WG co-ordinate on 
conformance requirements and the engineering of superior solutions to 
those which were implemented in HTML 4x/XHTML 1.0, but which have been 
dropped, deprecated or materially altered in the HTML5 draft...

gregory.

reference 1: "assistive technology"
 * http://en.wikipedia.org/wiki/Assistive_technology

reference 2: WAI ARIA Roadmap
 * http://www.w3.org/TR/wai-aria-roadmap/#abstract

reference 3: User Agent Accessibility Guidelines, 2.0
 * http://www.w3.org/TR/UAAG20/ (first public working draft)
 * http://www.w3.org/TR/UAAG10/ (UAAG, 1.0)

reference 4: NVDA (NonVisual Desktop Access)
 * http://www.nvda-project.org/
 * http://www.nvda-project.org/blog/

reference 4: Orca
 * http://www.gnome.org/projects/orca/index.html

reference 5: OATSoft (Open Source Assistive Technology)
 * http://www.oatsoft.org/
 * http://www.oatsoft.org/Info
 * http://www.oatsoft.org/Forge
 * http://www.oatsoft.org/Software

reference 6: FireVox
 * http://firevox.clcworld.net/features.html

reference 7: Opera Voice:
 * http://dev.opera.com/articles/voice/

----------------------------------------------
Theory helps us to bear our ignorance of fact.
    -- George Santayana, "The Sense of Beauty"
----------------------------------------------
Gregory J. Rosmaita: oedipus@hicom.net
Camera Obscura: http://www.hicom.net/~oedipus/
----------------------------------------------

---------- Original Message -----------
From: Dan Connolly <connolly@w3.org>
To: "Gregory J. Rosmaita" <oedipus@hicom.net>
Cc: Al Gilman <Alfred.S.Gilman@IEEE.org>, "Michael(tm) Smith" 
<mike@w3.org>, Chris Wilson <Chris.Wilson@microsoft.com>, HTML WG <public-
html@w3.org>, W3C WAI-XTECH <wai-xtech@w3.org>
Sent: Wed, 09 Apr 2008 18:09:26 -0500
Subject: Re: Discussing ARIA in HTML5 integration

> On Wed, 2008-04-09 at 23:45 +0100, Gregory J. Rosmaita wrote:
> [...]
> > and, what i said at last week's HTML WG telecon was that ARIA 1.0 is 
> > needed today and yesterday, and that embedding ARIA in HTML5 test 
> > cases is a non sequitur, because:
> 
> Ah... that's what you were responding to. When I asked on the
> phone, you said you weren't responding to anything I had said.
> I was at a loss.
> 
> We need test cases for every feature of HTML 5;
> I prefer to use test cases whenever we make a
> technical decision as a WG. So I'm confused when
> you say that test cases for ARIA in HTML 5
> are a non sequitur.
> 
> -- 
> Dan Connolly, W3C http://www.w3.org/People/Connolly/
> gpg D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E
------- End of Original Message -------

Received on Thursday, 10 April 2008 00:50:14 UTC