Naming Form Controls with the Label Element recommends incorrect technique

[dotjay]

On the Providing Accessible Names and Descriptions practices page under the heading Naming Form Controls with the Label Element, the text describes two techniques, commonly known to accessibility professionals as implicit and explicit labeling.

Implicit:

Wrapping the checkbox and the labeling text in a label element as follows gives the checkbox an accessible name.

Explicit:

A form control can also be associated with a label by using the for attribute on the label element.

The guidance goes on to state the following, which I believe to be incorrect:

When possible, use the above encapsulation technique for association instead of the following for attribute technique.

However, explicit labeling using for and id attributes has long been established as the preferable technique due to better support in browsers and assistive technology software, which is echoed by the WAI's own tutorial on Labeling Controls (with my emphasis):

Associating labels implicitly
In some situations, form controls cannot be labeled explicitly. For example, a content author might not know the id of a form field generated by a script, or that script might not add an id at all. In this case, the label element is used as a container for both the form control and the label text, so that the two are associated implicitly. Generally, explicit labels are better supported by assistive technology.

This is backed up by experts reporting issues with voice control software like Dragon Naturally Speaking not working with implicit labeling techniques, and more recent testing with screen readers.

I suggest that the advice on the Providing Accessible Names and Descriptions page be reworded to something like:

When possible, use the following for attribute technique for association instead of the above encapsulation technique.

[patrickhlauke]

The guidance goes on to state the following, which I believe to be incorrect

good lord, well spotted. that sounds like advice from some long-forgotten past (and then it turns out this was all because of a bug in Netscape Navigator 4.x on OS 9, or something)

[patrickhlauke]

Pull request submitted here @dotjay #2871

[daniel-montalvo]

Thanks much @dotjay and @patrickhlauke for your comments and for opening a PR on this.

@mcking65 Could we discuss this in the next APG meeting?

[mcking65]

PR #2882, which resolves this issue is merged. The change is scheduled for publication tomorrow, December 19.

Thank you @dotjay and @patrickhlauke for raising the issue and shaping the resolution.