W3C home > Mailing lists > Public > w3c-wai-ua@w3.org > January to March 2012

Re: Feasibiity as a factor for Level Selection

From: Charles McCathieNevile <chaals@opera.com>
Date: Mon, 30 Jan 2012 08:43:27 +0100
To: "Jeanne Spellman" <jeanne@w3.org>, "Jim Allan" <jimallan@tsbvi.edu>
Cc: wed@csulb.edu, "WAI-UA list" <w3c-wai-ua@w3.org>
Message-ID: <op.v8vzqpm7wxe0ny@widsith.eng.oslo.osa>
On Fri, 27 Jan 2012 21:40:45 +0100, Jim Allan <jimallan@tsbvi.edu> wrote:

> Going by Wayne's original document I would say the amount or quality
> AI involved simple guess versus complex heuristic..
> I don;t thing we can say an existing implementation indicates ease,
> only that there was a good enough business case to make a feature.

It's really a question of whether the technology is available and  
reliable. If someone can implement it easily, then it is probably not  
terribly hard - although it might be hard to put into a specific product.

The problem for UAAG in particular and accessibility in general is that it  
is too easy for developers to say "but this is really hard". There was a  
specific exclusion to 508 saying you didn't need to do things like make  
selection of brush types in a painting program keyboard controllable.  
Having contributed to an entirely keyboard-controlled program for  
painting, which allowed multiple brush selection, it is pretty clear this  
was the result of a self-interested company lobbying for a lighter  
requirement, rather than any reality-based exclusion because something is  
too difficult.

Which is why I like Wayne's approach. Indeed, deterministic vs inferential  
might be a very large part of the question of ease - since the former  
implies programming against a known target and the latter asks developers  
to imagine some algorithm that cannot be described.



> On Fri, Jan 27, 2012 at 12:45 PM, Jeanne Spellman <jeanne@w3.org> wrote:
>> Agreed, this is a useful analysis and I have already grabbed part of it  
>> to put into the proposal for the level definitions.
>> My question is: deterministic vs. inferential is one part of ease, but  
>> it certainly isn't all of it.  What are the factors that define ease?
>> We already have existing implementations indicate ease, but what else?
>> What else makes a feature hard to implement?
>> jeanne
>> On 1/27/2012 1:10 PM, Jim Allan wrote:
>>> Wayne,
>>> thanks for doing the thinking and writing on this. A good place to
>>> start. I agree with Chaals it strikes a good balance for ease vs
>>> impact. That's where we as a group can decide to push the level up or
>>> down.
>>> Jim
>>> On Thu, Jan 26, 2012 at 7:40 PM, Wayne Dick<wayneedick@gmail.com>  
>>>  wrote:
>>>> Hi All,
>>>> I wrote this to help us narrow feasibility to a reasonable range, so
>>>> we could use it in choosing Level A through AAA.  We don't want to let
>>>> developer's off reasonable tasks, but we can't require excessive
>>>> development.
>>>> http://www.csulb.edu/~wed/Feasibility.html
>>>> Wayne

Charles 'chaals' McCathieNevile  Opera Software, Standards Group
     je parle français -- hablo español -- jeg kan litt norsk
http://my.opera.com/chaals       Try Opera: http://www.opera.com
Received on Monday, 30 January 2012 07:44:14 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:49:41 UTC