Re: XTECH: ARIA in CSS for generated content (Was: CSSWG: Need fallback alternatives for CSS generated *text* content containing unicode glyphs)

On Feb 28, 2013, at 11:52 PM, "Schnabel, Stefan" <> wrote:

> So, the general idea seems to add extra ARIA info into CSS to control better the screen reader output.

Not limited to screen readers, but that's the primary case, yes.

> And this will have probably performance impacts, if I understand that right.

This is not necessarily the case, but it's possible.

> Another valid alternative would be IMHO to sit together with the screen reader guys and give them a general interpretation speech pattern
> what they have to announce in the use cases described. For Apple, this should be a home run. Others may follow or may not.

I'm not sure what you mean. WebKit supports some aspects of CSS 3 Speech and Web Speech API already, if that's what you're referring to.

This idea about exposing accessibility properties of CSS-generated content like pseudo elements. To my understanding, this is not specifically related to speech, but is more to expose the right thing to accessibility APIs, which can be accessed by other tools, including screen readers that use speech, but also automation frameworks, and other assistive technology.

> Anyway, the effort on screen reader side to implement these may be comparable to the effort implementing the proposed extensions for ARIA 2.0. In addition, who grants that even after implementation, speech output will not differ again from what you intend?

I dont understand this question. I believe this would result in rendering engine changes only. No screen reader change should be necessary.


Received on Friday, 1 March 2013 23:48:11 UTC