failure for not identify required fields and input format before submit

[goetsu]

I think we need a failure to cover the case where no information is given to identify required fields and input format before submit because for me that a clear violation of SC 3.3.2

[goetsu]

example :

Description

This failure ensures that people with visual, motor or cognitive disabilities will recognize required fields and input format before to submit the form to minimize input errors. Fields that are required or that need specific input format are frequently formatted in fixed, distinctive ways, and authors may feel that just providing visual formatting of the fields will be sufficient to identify them. However, this information must be be available programatically and visually and it must be available before submitting the form.

Examples

a form with three fields of which one required field and one with a specific format

Failure example 1 : no information on required field and specific input format:

<p>
     <label for="name">Name</label>
     <input type="text" id="name">
</p>
<p>
     <label for="city">City</label>
     <input type="text" id="city">
</p>
<p>
     <label for="date">Date of birth</label>
     <input type="text" id="date">
</p>
</form>

Failure example 2 : visual information only

<form>
<p>Field with * are mandatory</p>
<p>
     <label for="name">Name</label>
     <input type="text" id="name">*
</p>
<p>
     <label for="city">City</label>
     <input type="text" id="city">
</p>
<p>
     <label for="date">Date of birth</label>
     <input type="text" id="date">(yyyy/mm/dd)
</p>
</form>

Failure example 3 : programatically associated information only

<form>
<p>
     <label for="name">Name</label>
     <input type="text" aria-required="true" id="name">
</p>
<p>
     <label for="city">City</label>
     <input type="text" id="city">
</p>
<p>
     <label for="date">Date of birth</label>
     <input type="text" id="date" aria-label="Date of birth (yyyy/mm/dd)">
</p>
</form>

Tests

Procedure

• try to submit the form without providing any information, if error message appear to signal required fields check that this information was available before submitting the form in the accessible name programatically associated with the fields and with a visible text or cue positioned near the fields.
• try to submit the form after filling text input and textarea with "test" content, if error message appear to signal wrong input format check that this information was available before submitting the form in the accessible name programatically associated with the fields and with a visible text or cue positioned near the fields.

Expected Results

If check #1 or check #2 are false, then this failure condition applies and the content fails this success criterion.

[jnurthen]

Why is example #2 a failure?

[awkawk]

I'd say that example 1 fails 1.3.1 but not 3.3.2, example 2 fails 3.3.2 but not 1.3.1, and example 3 fails both.

[jnurthen]

Why does #2 fail? The default rendering of required on many browsers includes an indication that the field is required. In these cases it is fine for 3.3.2 IMO

[awkawk]

Ah, right. Assuming all of the necessary accessibility support then #2 would be fine for 3.3.2.

[goetsu]

Ok then change replace required with aria-required in example 2 if you prefer

[goetsu]

Or a class="required" and a JavaScript associated with that classname to make the field required.

My point with that failure proposal is to make more clear that not mentioning required status of fields in an accessible manner (visually and programmatically) PRIOR to submit is a failure.

[jnurthen]

I'm fine with either of those so long as it is made clear in the failure example that there is not CSS indicating that the field is required based on that attribute.

The test procedure also needs changing to be Check#1 AND Check #2 are false then failure condition applies, not an OR.

[goetsu]

Update of the proposal to include specific format.

[jnurthen]

We need to ensure that this does not fail content where this information is not known ahead of time. There are instances where fields are required only based on complex business logic which is ONLY available on the server following a form submit. This is the same for everyone so don't see how this could possibly be an accessibility violation.

[goetsu]

i'm not against this exception if you think it's needed but be careful because you can apply "the same for everyone" argument even for presence of label or the whole SC 3.3.

SC 3.3 exist in WCAG to help user to avoid mistake because people with disabilities are more likely than others to make mistakes

[joshueoconnor]

I like this draft - good work.