RE: CSS flexbox

I would like to emphasize on the point that the W3C specification itself use already inserts disallowing for a text scan on the page for the requested word.

Example: http://www.w3.org/TR/css3-flexbox/

Searching for the word "Example" (to move "by example" through the spec) yields nothing since example sections are based on an official (?) W3C template that is based on injections.

<div class="example">

.example:before {
    content: "Example "

As Rich said, the cat is already out of the box and user agents + AT need to arrange with this in near future.

Best Regards
Stefan

From: James Craig [mailto:jcraig@apple.com]
Sent: Mittwoch, 8. Oktober 2014 01:39
To: Richard Schwerdtfeger
Cc: Janina Sajka <janina@rednote.net(janina@rednote.net); Michael Cooper; W3C WAI Protocols & Formats; Ted O'Connor
Subject: Re: CSS flexbox

On Oct 7, 2014, at 3:44 PM, Richard Schwerdtfeger <schwer@us.ibm.com<mailto:schwer@us.ibm.com>> wrote:
Janina,

We would benefit from a meeting with CSS at TPAC. These are the topic areas:

1. The injection of content into HTML

We went a bit deeper into name computation pertaining to CSS in the ARIA implementation meeting today.

The new content property injects content into a web page. ... Looking at section 12.2 (http://www.w3.org/TR/CSS21/generate.html) you can use CSS to inject a broad range of attributes that an accessibility test tool would need to parse from computed style. That is a lot to ask of accessibility test tools. Joseph, Cynthia, and I were not favor of CSS injecting all this content but the cat is already out of the bag.
Do you have a specific agenda? Some of us have already provided feedback on generated content that has resulted in updates to the "content" property, list counters, and a new "alt" property in CSS4.

The CSS4 draft is not yet available. Here's the current URL for CSS3. http://dev.w3.org/csswg/css-content/
2. The impact of CSS Flexbox on navigation order. Flexbox changes the visual order and should logical navigation order. So, if we follow tabindex alone I think we are going to have a problem for authors. What would be good would be to be able to direct navigation based on flow vs. the DOM structure. Either this or we have to have browser implement some sort of flowto into the a11y api mapping to enable ATs in screen navigation using a logical order.
Again, I would recommend writing a specific agenda if you want to utilize joint meeting time. This is another area where some members of the CSS WG have provided accessibility feedback previously. While incomplete, most of the necessary flow information can be detected using the following properties:

- flex-direction
- flex-wrap
- flex-flow
- writing-mode
- order

There's even a "Reordering and Accessibility" section in the document.
http://dev.w3.org/csswg/css-flexbox/#order-accessibility

Outside of that spec, there is related ongoing work in the CSS-UI module for navindex, which solves some of the index problems that may be caused by flexbox. Navindex is woefully incomplete, but PF should not propose with CSS WG unless there is some specific feedback to be given.

http://dev.w3.org/csswg/css-ui/#nav-index0

James