3.3.2 labels or instructions: additional info in placeholder

[detlevhfischer]

We are discussing whether additional information supplied with text field labels (such as date format) needs to be permanently visible to meet 3.3.2, or if cases where a descriptive label such as "Date of birth" is permanently visible but additional information is provided as placeholder text (e.g., placeholder="DD/MM/YYYY") would minimally pass 3.3.2. Any opinions?
See BIK-BITV/BIK-Web-Test#499

[TestPartners]

I reluctantly conclude that the additional information is not mandatory. The Understanding page says "The intent of this success criterion is to have ... instructions or labels that identify the controls in a form so that users know what input data is expected. ... Instructions or labels may also specify data formats for data entry fields." The word "may" is key.

Arguably, people with disabilities are not at a greater disadvantage than anyone else if the instructions are not persistent. On the other hand, people with short-term memory difficulties (and I include myself in this group) or other cognitive impairments might well be at a greater disadvantage. However, this is not relevant if the purpose of the SC is to identify the purpose of the form control.

The required data format could be presented in an error message, so it does not necessarily have to be visible initially even though that would be a better user experience.

[mbgower]

There are now a number of separate issues open about using placeholders, from several different vantage points (several opened by you!):

• 1.4.11 Non-text contrast understanding doc references placeholder as a passing technique, despite it being a transient state #4413
• 1.4.3 placeholder contrast requirement if placeholder text is redundant #4343
• 3.3.2: provide failure examples/techniques for placeholder and <select> relying on default <option> only #2509
• Does 2.5.3 "Label in Name" apply to placeholder (where used as only visible label)? #1700
• Text in Understanding 3.3.2 allows something functionally identical to HTML placeholder #1136
• Applicability of 2.5.3 Label in Name when no label is present (or just placeholder) #855

Taken as a whole, they suggest that placeholders pose many challenges to different users. Perhaps it's time the TF undertook something more holistic on placeholder guidance.