WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: WCAG 2.1 - 1.3.5 - How to capture a violation?

for

From: Jared Smith
Date: Jul 25, 2018 11:15AM

Patrick wrote:

> This has been the crux of the problem around 1.3.5. It really means "machine-readable", i.e. it needs to be part of a defined taxinomy that tools can use to determine the purpose of a control (and potentially offer useful behaviour, such as autofilling form fields).

Jon wrote:

> From what I understand the goal of the criterion was to ensure that there was a machine readable way to know exactly what the purpose of the field was

I cannot glean a semblance of this interpretation from the normative
text. The SC simply does not indicate that the field must align with a
defined or machine-readable taxonomy. It instead states that WHEN the
field aligns with the defined taxonomy (the autofill names) that the
purpose be programmatically determinable. There's a big difference
between what you are stating the goal of the SC is and what it
actually says.

As an example, <label>Name: <input type="text"></label> clearly aligns
with an autofill value - it collects the user's name. The SC therefore
requires that its purpose be programmatically determinable - which it
clearly is because of the associated label of "Name". (Interestingly,
this would also work in any useful form filler, but that's rather
beside the point.)

If this code is not sufficient, then can you provide an explanation as
to how it fails to meet the normative text of 1.3.5? It's quite
possible I'm misunderstanding something here.

> I'd suggest that you log an issue on the github repository for WCAG with what you think would help clarify this in the understanding documents of WCAG 2.1

There's not anything wrong with the existing technique. It's certainly
one way to meet the success criterion. But if the intent of the
success criterion is that conveying the purpose via an accessible name
is not sufficient and/or that the purpose must be conveyed via a
defined taxonomy, then I don't think this can be addressed without
modifying the success criterion itself to indicate this.

Thanks,

Jared