Re: RE: working definition of baseline

On Sun, May 08, 2005 at 08:49:44PM -0500, John M Slatin wrote:
> 
> I have a feeling that this goes back to something that Wendy (I think)
> said in an earlier thread about baseline and 4.2-- I'm remembering a
> distinction drawn between technologies *required* by the content and
> technologies *used* by the content.  
> 
> The idea was that in the absence of a *required* technology content
> would not be perceivable, operable, understandable, and robust, and
> would not degrade gracefully-- it would simply break. The content
> *would* degrade gracefully in the absence of technology that was merely
> "used."

This is the right distinction to be making here.

If the baseline is defined as those technologies which are *required* in 
order for the content to be perceived/operated *at all* by any user 
agent, then that's very different from defining it as those technologies 
which the author thinks desirable to assume in evaluating conformance.

John asked for an example. Suppose there is a metadata format (Lisa has 
defined one) in which text alternatives can be associated with non-text 
content. Assume for the sake of argument that there are some 
UAAG-conformant, or largely UAAG-conformant user agents which support 
this format, and I publish a Web site in which all the text alternatives 
(along with other information required by WCAG) are provided in this 
metadata format.

Now I publish my baseline. My content will function in a user agent that 
doesn't support the metadata format, but it will only satisfy guideline 
1.1 under the assumption of a user agent that does, due to the technique 
I have chosen to implement the success criterion.

If my baseline is required to be the minimal set of technologies needed 
for the content to be rendered at all, then I have failed the success 
criterion because the metadata format isn't included in the baseline as 
it degrades gracefully with respect to user agents that don't support it 
- they just won't get the text alternatives. If I can define my baseline 
more broadly then I can pass guideline 1.1.

Suppose we opt for the first approach - my baseline is defined in terms 
of the absolutely essential technologies needed for the content to be 
rendered, which doesn't include my metadata format. Now, I can handle 
this by using "browser sniffing" techniques (HTTP user agent headers or 
CC/PP) to deny access to my content to user agents that don't support 
the metadata format. I've just made it into a "required" technology for 
my content by shutting out user agents that don't support it. Having 
done this, it appears that I can now claim guideline 1.1 conformance, 
which presumably is not the result we want.

Thoughts?

Received on Monday, 9 May 2005 03:20:33 UTC