> if the content author makes a change to one aspect of the control, and that change visually impacts other aspects of the same control, then the author is responsible for all of her actions, not just the one.
In which case, can you point to a link or button on a live (non-test case) site where the exception would apply? Unless I’m missing something, if that is the interpretation then there’s no point in the exception. For example, changing the background colour of the page (even just to white) would mean all components are then in scope.
> We cannot accept that Firefox's default back dotted lines, situated on a CSS-modified background color of Navy blue, is acceptable simply because "that's the default focus indication".
For me (personally), it is to do with where we draw the line on user-agent vs web content. My current focus indicator is 9px thick with three colours. (Just because I’ve been experimenting.)
[cid:image003.png@01D3FC65.53EACE70]
More generally, how does an author know what a particular user-agent will use?
* Chrome & safari on Mac use multiple shades of light blue,
* Chrome on windows uses a darker (medium) blue,
* Firefox is dark grey dots (I think, I’ve over-ridden it),
* Edge uses black dots on a white line.
As a web-content guideline, if we have an exception for default UA behaviour (which we do), we can’t vary it based on the user-agent used. Therefore, if the author doesn’t override the UA focus indicator, we can’t pass/fail based on the background colour compared to light blue, medium blue, and black & white at the same time.
-Alastair
