Re: Action-1634: the switch element and <title> and <desc>

When an author uses aria-describedby it is considered an override of desc
as the description. The aria-describedby may point to text that is visible.

We would definitely not want to concatenate the title and the desc
together. They have two distinct purposes (title is a short name and desc
is a longer description,usually to provide more detailed information about
the object and it can be quite long).

Rich

Rich Schwerdtfeger



From:	Amelia Bellamy-Royds <amelia.bellamy.royds@gmail.com>
To:	"White, Jason J" <jjwhite@ets.org>, "public-svg-a11y@w3.org"
            <public-svg-a11y@w3.org>
Cc:	Richard Schwerdtfeger/Austin/IBM@IBMUS
Date:	05/17/2015 06:07 PM
Subject:	Re: Action-1634: the switch element and <title> and <desc>



The way I wrote it, if both <title> and <desc> are provided, plus an
aria-label attribute, the title tooltip text would basically be discarded.
I think this is consistent with how the HTML mappings spec uses the title
attribute.

Ideally, this would be because the author knew what they were doing and had
a short tooltip for visual users and a more informative label + description
for AT that included the same information with more complete context.  But,
it is still throwing out information that is available to screen + mouse
users.

If there is also an aria-describedby attribute then both the <title> and
the <desc> are thrown out unless the describedby attribute explicitly
refers back to the same element.  I can't think of a case when this would
be intentional (since <desc> does not have any non-accessibility
functionality).  The most likely reason is an author trying to use both
visible and non-visible text to provide a complete description, and not
realizing that one replaces the other.

I kind of like the idea of concatenating the <title> content onto the end
of the <desc>, but I think it would probably need to be distinguished from
the main description.  Ideally, it would be clearly announced as a
"tooltip" or "mouseover text" for the content.

Does anyone know if any accessibility technologies do this with HTML title
attributes?  Of course, one difference between SVG and HTML is that the
HTML title would still be available to ATs as an attribute on the DOM
object, even if it wasn't mapped to a specific property in the
accessibility API.  Because the SVG title is a separate element, and that
element is not itself represented in the accessibility tree, I'm not sure
if ATs would have any way to access it.

Similarly, I'm not sure whether there is any way to make sure that <desc>
content is always available as additional text.

~ABR

On 17 May 2015 at 16:20, White, Jason J <jjwhite@ets.org> wrote:

  > On May 17, 2015, at 14:44, Amelia Bellamy-Royds <
  amelia.bellamy.royds@gmail.com> wrote:
  >
  > Second thoughts on title and desc:
  >
  > In HTML, the title attribute can be used to create the accessible
  description if it is not used as the accessible name.  We should
  explicitly define the same behavior for SVG.
  >
  > In other words, if the author specifies aria-label or aria-labelledby,
  use that for the name, but if they also include a <title> element (as a
  tooltip for visual users), that should be used as the description if no
  other description is provided.  This would support the common practice in
  data visualizations of using title tooltips to provide additional
  information, such as the numerical data.
  And if the author also provides a DESC element, should its contents be
  appended to the description?

  I know very well that supplying aria-label/aria-labelledby + title + desc
  is ill-advised, at the very least. However, it would seem reasonable
  still to treat the ARIA attribute as overriding the accessible name and
  to concatenate title and desc to form the accessible description.

  Are there any other edge cases to consider?