Re: SC 1.4.11

John Foliot wrote:

> I'm saying we use the precise reading of the exception clause



OK, going down that route: A state [1] is defined as a characteristic of the component, therefore is part of the component.



Therefore you could read that as:
“the appearance of [any aspect of] the component is determined by the user agent”.



Or, you could read that as:

“the appearance of [that aspect of] the component is determined by the user agent”





> I am at a loss as to why we don't want to pursue that approach.



The literal interpretation (any aspect of a component) will be counterproductive as it leads to absurd situations:

  *   A change of font, size, or pretty much any CSS attribute (for “appearance”) means the exception doesn’t apply.
E.g. a user-agent such as a TV fixes the focus-style but not font-styles, that is what the exception was intended for, but it wouldn’t apply.

  *   A change to the background of the component means it doesn’t apply (e.g. a basic page which has the background and foreground colours set, but not link pseudo-styles).

  *   For components (like checkboxes) which have limited styling options, that would lead to either odd design choices (white patches in a form!?), or devs replacing default components with custom ones (impacting other groups).



I think the “that aspect of” interpretation provides enough flexibility to be useful.



Overall, I think we can agree that:

  *   We should advocate for best practice with good contrast (in the understanding doc and elsewhere);
  *   If the author provides the indicator it is in scope.
  *   The use of default focus indicators is not that common.



The disagreement comes from how we assume people will interpret the language for the exception, and who’s responsible for poor default-focus indictors.



Cheers,



-Alastair