Re: UNS: Re: WCAG considering amending F65 to NOT fail missing ALT text if title or aria-label is present

David wrote:

> This statement actually demonstrates a lack of distinction between a 
> validation error and a WCAG error, which concerns me because if an 
> accessibility expert misses this distinction, which James Craig pointed 
> out in the first post on this issue, what can we expect from Johnny 
> lunch box developers?

Indeed, SC 4.1.1 itself mixes the two concepts. A page can hava 
duplicate IDs and/or duplicate attributes and/or nesting errors 
according to the specs, and still be completely accessible. Many parsing 
errors in the Real World have no influence on accessibility, although of 
course badly constructed code can sometimes have accessibility implications.

> WCAG techniques are all 
> about what works today rather than what should work... and it’s always 
> been that way, because if the technique is not accessibility supported 
> today ... it will not work... the place for experimentation, and 
> innovation is not in WCAG techniques... WCAG is about what works today 
> for accessibility and this how we make “web content more accessible” so 
> naturally we will follow trends in our techniques.

This is another big debate that in my opinion is unsolved...

Almost all PDF techniques are only accessibility supported on Windows 
platforms and using Adobe Reader, and there is no statement in the 
techniques reflecting this behaviour. This is a real big issue for many 
blind users that cannot properly access PDF documents if they are not 
using Windows. Some of the Flash techniques just do not work as they are 
described, and probably several other techniques have partial support 
without mentioning it.

My experience with developers is that they read the sufficient 
techniques as "working under all possible circumstances", since most of 
them have no accurate information (or no information at all) about their 
accessibility support, thus ignores Conformance Requirement #4.

On the other hand, some people read failures as "they always lead to 
failure of conformance", ignoring that the failing content could be "not 
relied upon", and therefore it could conform meeting Conformance 
Requirement #5.

Moreover, Conformance Requirement #1 teorethically allows a page to be 
completely inaccessible but still conformant, provided that there is an 
alternate version that can be reached through an "accessibility 
supported mechanism". But it does not specify where this mechanism must 
be placed, so in theory it could be the last link of the webpage, making 
it completely unusable by real users with disabilities.

So in my view the accessibility support is so poorly documented in the 
official WCAG documents that it leads to either too strict or too loose 
interpretations, depending on who is affected by the issue. Developers 
tend to read sufficient techniques as "always valid" and failures as 
"always failing". Meamwhile, users will suffer the consequences of 
not-always-sufficient techniques and simply ignore the not-always-failures.


Received on Tuesday, 26 November 2013 11:29:33 UTC