remove reference to whitespace #135
Conversation
resolves #128 Per https://www.w3.org/2022/08/11-aria-minutes#t05, the WG resolved that the empty string should cause the attribute to be ignored, but not whitespace. So, reference to whitespace can be removed from this spec. However, referring to whitespace will remain in the ARIA spec and THERE it will link to the ascii-whitespace definition where applicable. Additionally, rather than a user agent requirement, it will become an authors requirement on not specifying whitespace as the value of an aria-roledescription
|
@jcsteh can you please review this. |
There was a problem hiding this comment.
One question, I'll do what ever we decide here for aria-brailleroledescription: https://github.com/w3c/core-aam/pull/134/files
|
From reading the IRC log, I'm somewhat surprised the WG agreed to allow authors to suppress the role description, but I acknowledge Aaron made a good point that the more complicated the logic is, the slower the implementations will be, and the more confused the authors will be. This is simpler, and allows a legitimate technique that is sometimes necessary (even if it can also be a footgun). |
|
My understanding of this is that First, why do we explicitly handle an empty attribute at all? Is this just "easy repair" for authoring error (as opposed to checking for white space, which would complicate the browser implementations)? Second, @cookiecrook's comment makes me worry that I'm misunderstanding something here:
What do you mean by "suppress the role description" here? Are you referring to |
|
@jcsteh "suppress the role description" is in reference to Scott's comment at the top, and from digging through the linked minutes:
I was not in that conversation, so I may be mistaken, but if I understand correctly, you are correct that |
There was a problem hiding this comment.
Thanks @cookiecrook. My understanding appears to be correct then.
This looks fine to me. I'm not sure how worthwhile it is to explicitly support the empty string; it seems like a very obvious and easy to explain authoring error. That said, it's easy to implement and I have no real objection to it.
SHA: fbad1bd Reason: push, by @spectranaut Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
resolves #128
Per https://www.w3.org/2022/08/11-aria-minutes#t05, the WG resolved that the empty string should cause the attribute to be ignored, but not whitespace. So, reference to whitespace can be removed from this spec. However, referring to whitespace will remain in the ARIA spec and THERE it will link to the ascii-whitespace definition where applicable. Additionally, rather than a user agent requirement, it will become an authors requirement on not specifying whitespace as the value of an aria-roledescription
Preview | Diff