Re: Effect of role=presentation on img elements with svg

True. It is true the image role was designed to group a collection of
images/content into a single image entity.

Steve Faulkner and I discussed this on SVG and Canvas and we agreed to have
both default to a group role. People don't normally think of an img as
containing descendants. ... Also, with minimal accessibility features added
to an SVG element (not the SVG element) I defaulted to a group role. The
<g> element in SVG is critial piece in SVG accessibility. In SVG we want to
have break graphics apart into meaningful elements of a drawing. With SVG2
they may be navigable with the keyboard and some pieces could event form
widgets (in the ARIA vernacular). So, for HTML 5.1 and SVG2 we defaulted
the SVG element to a role of <group> . The author has the ability to make
it an ARIA image by supplying the "image" role.


Rich Schwerdtfeger

From: Joseph Scheuhammer <>
To: Richard Schwerdtfeger/Austin/IBM@IBMUS, James Craig
Cc: Bryan Garaventa <>, Cynthia Shelly
            <>, Steve Faulkner
            <>, ""
            <>, Joanmarie Diggs <>,
            "" <>, "T.V Raman"
            <>, " WAI-PFWG"
            <>, "" <>,
            Leif Halvard Silli <>
Date: 03/12/2014 11:43 AM
Subject: Re: Effect of role=presentation on img elements with svg

Somewhat of a tangent, but may have some relevance.

On 02/03/2014 01:24 PM,  James Craig wrote:
> In the case of accessible SVG images on OS X, the image has a rendered
> subtree, and the current version of AXImage (the AX API mapping) does
> not support descendant contents, so it’s mapped, on that platform, to
> AXGroup (a platform-specific role that is not the same as the ARIA
> “group” role).

The ARIA img role is a container:  " A container for a collection of
elements that form an image." (

Both the IA2 GRAPHIC role and the ATK/AT-SPI  IMAGE role can have
accessible children.

  Given this construct:

<div role='img' ...> ... </div>,

FF creates an a11y sub-tree with the image accessible as the root, and
with accessible descendants depending on what's inside the <div>.  FF
could (maybe already does) map the svg <image> element the same way.


'A: After all, it isn't rocket science.'
'K: Right. It's merely computer science.'
              - J. D. Klaun -

Received on Sunday, 16 March 2014 12:27:48 UTC