Re: Finding agreement on common purpose

On 16/01/2018 12:27, Alastair Campbell wrote:
>> But wouldn't a conforming AT go looking for the exact token as defined
>> in HTML? If not, where are these mappings defined?
>      
> The mapping to a technology should be at the technique level, rather than in the “technology agnostic” guidelines doc.

I don't think there should be any mapping at all. There should be a 
simple set of "purposes" defined in WCAG.


> 
> So the logic is:
> - WCAG provides a list of purposes for inputs (and hopefully later links & buttons);

Agreed.

> - If you are using HTML, here is an HTML technique for applying those purposes to inputs, which maps to HTML autofil.

To the extent that the current SC language is too technology specific, I 
agree. I still don't think that mapping gives the AT what it needs to 
make the concrete determination of a qualifying control's purpose.

In HTML it would be conforming to use one of the HTML tokens, or a 
completely different token providing there was some mapping document 
somewhere that pointed out the relationship between the two. That seems 
to be a good way to make the SC useless to the people who'll benefit 
from the concrete determination of a control's purpose.



> 
> Ideally we’d refer to a specific spec for all of it, in a similar way to 4.1.2 being primarily defined by ARIA.  There is such a spec, but it is very wide-ranging and some aspects are further ahead than others.


> 
> Over the history of this SC it has been whittled down to practically nothing, with the HTML autofill being a very small toe in the door to get something in. However, I think the principle of how it works should continue to work if it is expanding in future.
> 
> Cheers,
> 
> -Alastair
> 

-- 
@LeonieWatson @tink@toot.cafe Carpe diem

Received on Tuesday, 16 January 2018 12:38:26 UTC