Re: aria-readonly and plain elements

Thanks Rich, I appreciate the clarification. As long as we have a spec that
everyone implements, it's for the best.

That said, our team finds it counterintuitive that aria-readonly can affect
an <input> or <textarea> but not a contenteditable. It seems inconsistent.
We can live with it but I thought I'd mention it because authors may also
be confused by that.

I'll take a closer look at the ARIA in HTML document.

Aaron

On Wed, Jul 26, 2017 at 7:51 AM Rich Schwerdtfeger <richschwer@gmail.com>
wrote:

> Hi Aaron,
>
> Restrictions on the use of ARIA in host languages is defined by the host
> language (HTML or SVG). There is a spec. called ARIA in HTML that would
> state this.
>
> Here is the spec.:
>
> https://www.w3.org/TR/html-aria/
>
> Which states:
>
> Do not set aria-readonly="true" on an element that has a contenteditable
> attribute set.
>
> The implicit host language semantics states that for content editable it
> is equivalent to aria-readonly=“false”
>
> This was released in March and the restriction on contenteditable was new
> at that time.
>
> This also should mean that aria-readonly=“true” is an authoring error and
> it should be ignored.
>
> In the HTML AAM (https://www.w3.org/TR/html-aam-1.0/) for content
> editable it states in a comment:
>
> If the element has the contenteditable attribute and aria-readonly="true",
> User Agents *MUST* expose only the contenteditable state.
>
> So, ignore aria-readonly=“true”.
>
> Rich
>
>
>
> Rich Schwerdtfeger
>
>
>
>
> On Jul 25, 2017, at 4:14 PM, Aaron Leventhal <aleventhal@google.com>
> wrote:
>
> I'm not sure I understand :/
>
> I could use some guidance as I'm touching this up in Chrome at the moment.
>
> Aaron
>
> On Tue, Jul 25, 2017 at 10:38 AM Rich Schwerdtfeger <richschwer@gmail.com>
> wrote:
>
>> That would need to be an HTML AAM restriction as HTML is the host
>> language and this would represent a conflict with the host language. Sent
>> from my iPhone. ARIA also applies to SVG.
>>
>> On Jul 24, 2017, at 7:33 PM, Aaron Leventhal <aleventhal@google.com>
>> wrote:
>>
>> To clarify, I'm talking about cases where there is no role.
>>
>> On Mon, Jul 24, 2017 at 7:19 PM Aaron Leventhal <aleventhal@google.com>
>> wrote:
>>
>>> It's my understanding that aria-readonly="true" should not be mapped by
>>> a user agent for <div contententeditable>.
>>> Also, aria-readonly="false" should not be mapped by a user agent for
>>> <div>
>>>
>>> Just checking that this is what the spec says, it's what's intended, and
>>> is considered desirable.
>>>
>>> Thank you,
>>> Aaron
>>>
>>
>