WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: WCAG 2.1 SC 1.3.5 Identify Input Purpose - TestingMethodology

for

From: John Foliot
Date: Sep 28, 2018 10:57AM

> The 'about the user' requirement being hard for software to determine
presents an interesting problem for plug-ins (or any other user agent)
consuming autocomplete to enhance presentation. They have no way of knowing
if the autocomplete attribute is 'about the user' or about something
else... Is this a problem, or am I missing something here?

Hi Mark,

You are essentially correct, and for the *functionality* assigned to
autocomplete, it is true that user-agents (browsers and password managers)
allow the end user to select appropriate stored data to complete the
fields. My current configuration(s) across multiple browsers actually
stores multiple 'personas' (i.e. I have a faked-out user: Mr. Joe. Q.
Tester, complete with a fake address, phone number and email address that I
use for testing, as well as my *actual* configurations - of which I have
both a Home profile with my personal email address, as well as a second
Work profile with my work address.)

However, I think the concern Jared is thinking of (a real concern) is that
it adds complexity to the testing, in that a tool alone cannot determine
"compliance" because that compliance is conditional on what type of data is
being collected (i.e. "of the specific user"). There are a number of other
use-cases that have surfaced that also add to "manual inspection
complexity" (imagine an online HR form that collects multiple "First
names"). As a representative of a company that also provides accessibility
compliance tools (Deque), we understand that every time you have to include
a more manual inspection, it adds to complexity, potential
"false-positives", and potentially additional cost/time to evaluate. While
not show-stoppers, they are real concerns, especially at large volume
(enterprise) scale. However, like many other SC in WCAG, at this time I
cannot see a why around the need for manual intervention in the test
process.

Not sure if this adds anything new to the discussion, but wanted to reply
to your thoughts.

JF

On Fri, Sep 28, 2018 at 6:52 AM, Mark Rogers < <EMAIL REMOVED> >
wrote:

> This makes it a bit difficult to test this automatically. It's
> difficult to surmise whether a form is collecting information about
> THE user vs. information about SOME OTHER user. This SC would only
> apply to the former.
>
> The 'about the user' requirement being hard for software to determine
> presents an interesting problem for plug-ins (or any other user agent)
> consuming autocomplete to enhance presentation. They have no way of knowing
> if the autocomplete attribute is 'about the user' or about something else.
>
> For example, an online tax form for non-profits might use
> autocomplete='organization' and autocomplete='street-address' to make it
> easier to complete. That seems like it could interfere with the input
> purpose autocomplete use outlined in SC 1.3.5 and cause problems with the
> WCAG Non-Interference requirement (e.g. if the AT overlays an icon
> indicating 'Your home address here' on the autocomplete='street-address'
> element.) But removing autocomplete from the form makes filling it harder
> for anyone who finds typing difficult due to Parkinson's or Arthritis.
>
> Chrome deals with ambiguity of the autocomplete tokens by presenting a
> dropdown containing alternatives for autocomplete fields - an
> autocomplete='street-address' field might show addresses for home, work and
> a relative because you ordered them a gift last birthday.
>
> Is this a problem, or am I missing something here?
>
> Best Regards
> Mark
>
> --
> Mark Rogers - <EMAIL REMOVED>
> PowerMapper Software Ltd - www.powermapper.com
> Registered in Scotland No 362274 Quartermile 2 Edinburgh EH3 9GL
>
>
>
>
> On 27/09/2018, 18:16, "WebAIM-Forum on behalf of Jared Smith" <
> <EMAIL REMOVED> on behalf of <EMAIL REMOVED> > wrote:
>
> Mark Rogers wrote:
>
> > Doesn't the failure only happen when the input field purpose matches
> one of the input purposes listed:
> > https://www.w3.org/TR/WCAG21/#input-purposes
>
> Yes, and only when collecting information directly from and about the
> user completing the form.
>
> This makes it a bit difficult to test this automatically. It's
> difficult to surmise whether a form is collecting information about
> THE user vs. information about SOME OTHER user. This SC would only
> apply to the former.
>
> > I don't think autocomplete='on' counts as a pass, because ‘on' is
> the default if autocomplete is not specified, and means user agents use
> heuristics to guess the input type which seems at odds with the SC.
>
> Correct. This would not provide the level of field purpose
> identification necessary for this success criterion.
>
> > I think having more documented successes and failures would help
> clarify this.
>
> Yep. I know the documentation for this is still in the works.
>
> For what it's worth, I agree with Patrick's assessment of the examples
> you provided.
>
> Jared
> > > > >
>
> > > > >



--
*​John Foliot* | Principal Accessibility Strategist
Deque Systems - Accessibility for Good
deque.com