- From: Joseph Scheuhammer <clown@alum.mit.edu>
- Date: Thu, 3 Sep 2015 10:57:34 -0400
- To: James Teh <jamie@nvaccess.org>, Alexander Surkov <surkov.alexander@gmail.com>
- Cc: W3C WAI Protocols & Formats <public-pfwg@w3.org>
Thanks James,
> I assume this change wouldn't prevent authors from using
> aria-labelledby to label "li" tags or role="listitem" elements?
> Otherwise, this would definitely break Twitter and similar cases.
Your assumption is correct. Using aria-labelledby or aria-label remains
an option. The change involves removing the requirement of an
accessible name for listitems, whereupon name-from-content is no longer
relevant.
> While slightly off-topic here, I see in the linked thread for action
> 1710 that this has also been requested for role="link". I don't
> understand this one even in principle, as links can certainly be seen
> as interactive.
Yes, removing the name requirement and name-from-contents from
role="link" is controversial, since, as you say, links are widgets and
interactive. However, the counterexample that doesn't work is where a
table is the content of an <a> element, something like:
<a href="somewhere><table> ... </table></a>.
If the table is large, then a name computed from the table content would
be meaningless. Then again, why would somebody make a large table a
link? Perhaps it is just an exaggeration to make a point.
My own feeling is:
- accessible name should remain a requirement for the link role, since
it is a widget.
- name-from-content is one way to calculate the name.
- when the content of the link is complex, authors must supply a name
(e.g., aria-label, etc).
The problem is when authors ignore the latter point.
--
;;;;joseph.
'Array(16).join("wat" - 1) + " Batman!"'
- G. Bernhardt -
Received on Thursday, 3 September 2015 14:58:07 UTC