RE: Thoughts on roles and conformance to SC 4.1.1 Parsing

Steve,  thank you for the feedback, in your example http://www.paciellogroup.com/blog/2014/10/notes-on-fixing-incorrect-table-structure-using-aria/ you use ARIA and HTML structures overlayed by using role=”presentation” but both structures are nested according to the specification.    So the role attribute has no effect on the HTML semantic meaning/structure only the ARIA structure – that is HTML validation does not consider the role in determining semantic structure.

So in my example, while not optimal, and not following the first rule of ARIA it would be more conformant to just use ARIA roles and a div as this is more semantically correct in both the HTML and ARIA specifications.
<div role=”list”>
  <p role=”listitem”>Apples</p>
  <p role=”listitem”>Oranges</p>
  <p role=”listitem”>Bananas</p>
</div>
Jonathan

--
Jonathan Avila
Chief Accessibility Officer
SSB BART Group
jon.avila@ssbbartgroup.com<mailto:jon.avila@ssbbartgroup.com>
Phone 703.637.8957
Follow us: Facebook<http://www.facebook.com/#!/ssbbartgroup> | Twitter<http://twitter.com/#!/SSBBARTGroup> | LinkedIn<http://www.linkedin.com/company/355266?trk=tyah> | Blog<http://www.ssbbartgroup.com/blog> | Newsletter<http://eepurl.com/O5DP>

From: Steve Faulkner [mailto:faulkner.steve@gmail.com]
Sent: Friday, August 21, 2015 1:23 PM
To: Jonathan Avila
Cc: GLWAI Guidelines WG org; Michael[tm] Smith
Subject: Re: Thoughts on roles and conformance to SC 4.1.1 Parsing

Hi Jonathon,

these sorts of markup patterns are HTML conformance errors, and for good reason,  It is not a good idea period to use HTML elements contrary to the rules of the HTML content models

http://www.w3.org/TR/html5/introduction.html#restrictions-on-content-models-and-on-attribute-values


--

Regards

SteveF
Current Standards Work @W3C<http://www.paciellogroup.com/blog/2015/03/current-standards-work-at-w3c/>

On 21 August 2015 at 17:40, Jonathan Avila <jon.avila@ssbbartgroup.com<mailto:jon.avila@ssbbartgroup.com>> wrote:
ARIA roles allow for flexibility in particular to correct semantic structure when it is incorrect and not easily changeable.  I wanted to get others opinions on situations where ARIA roles are mixed with semantic structures and the HTML nu validator indicates missing nesting structures.  Accessibility support is also a related concern.

For example, we have an unordered list with child elements that are not list items.  We then add an ARIA role to the list items.  We might end up with something like the below:
<ul>
  <p role=”listitem”>Apples</p>
  <p role=”listitem”>Oranges</p>
  <p role=”listitem”>Bananas</p>
</ul>

The opposite might also be true – say we have li elements within a div with role=”list”.  A similar situation often occurs with grid/table structures that have some table semantics but then use divs with “gridcell”.  So in regards to SC 4.1.1 what are your thoughts on mixing native structures with ARIA roles?  Would it be best then to reapply the default native semantic role?

Jonathan

--
Jonathan Avila
Chief Accessibility Officer
SSB BART Group
jon.avila@ssbbartgroup.com<mailto:jon.avila@ssbbartgroup.com>
Phone 703.637.8957
Follow us: Facebook<http://www.facebook.com/#!/ssbbartgroup> | Twitter<http://twitter.com/#!/SSBBARTGroup> | LinkedIn<http://www.linkedin.com/company/355266?trk=tyah> | Blog<http://www.ssbbartgroup.com/blog> | Newsletter<http://eepurl.com/O5DP>