Re: ARIA roles for graphics

I think if we can define figure well enough that won't be a problem as
there is a corresponding HTML element. iconbutton will be tougher as most
platforms just treat this as a button. Icon in IA2 simply defines a small
fixed size picture. IA2 is implemented by FF and Chrome.

IA2
                                                                             
 IA2_ROLE_ICON    A small fixed size picture, typically used to decorate     
                  components.                                                
                                                                             



We went round and round on a lot of this with the DPUB ARIA spec. It was
determined that it would be best to put the taxonomy vocabulary precursor
at the front and minimize the amount of roles that went into the main ARIA
spec. One big reason for this is browser vendors, like Microsoft, wanted to
claim they were ARIA compliant. If we all of a sudden add a new role to the
main ARIA vocabulary then they fall out of compliant. Later on, when they
are ready, the can say they are ARIA graphics compliant or ARIA DPUB
compliant. ARIA has gained international recognition and a essential
standard to support accessibility. The browser manufacturers don't want to
lose their compliance claims. That does not prevent us from tweaking the
vocabulary a bit for things like role="img" in the ARIA spec. if it does
not break other things.

Rich


Rich Schwerdtfeger



From:	Amelia Bellamy-Royds <amelia.bellamy.royds@gmail.com>
To:	Richard Schwerdtfeger/Austin/IBM@IBMUS
Cc:	"public-svg-a11y@w3.org" <public-svg-a11y@w3.org>
Date:	06/25/2015 09:05 AM
Subject:	Re: ARIA roles for graphics



Regarding the figure example, I agree that this might not be the best
example.  I really wanted to emphasize the idea that part of an <svg> could
be it's own labelled figure, and that figures could be nested.  However, I
was also trying to keep it simple.  The result was maybe too simple, since
as you say each "sub-figure" is just a single element and should have
role=img.  I'll try to think up something better.

I recognize the problem with having two specs being developed in parallel
with different definitions for the same role.

I'm starting to think that it would be easiest if we can convince the main
ARIA team to pull the general purpose roles into the main spec: figure,
iconbutton, symbol, and the revised img definition.

We could keep graph-document and graph-object as the base elements in our
new graphical role taxonomy, which would also include more specific roles
such as graph-map, graph-chart, graph-axis, graph-legend, etc.

ABR

On 24 June 2015 at 14:49, Richard Schwerdtfeger <schwer@us.ibm.com> wrote:
  Thanks for the great work.

  Just a comment on the current draft. We wanted Figure to subclass region,
  yet the example for figure places "figure" role elements inside the HTML5
  <figure> element. These are really symbols inside a figure. If we want to
  apply an SVG role of "figure" we should do it on the the SVG element here
  and make it encompass symbols.

  Also, we should not have an "img" role in our spec. as it is in the ARIA
  1.1 spec. However, if we feel the the ARIA 1.1 spec. text for "img" is
  inadequate then we should raise an issue on the ARIA 1.1 spec. An intent
  of mine is to try and synch ARIA 1.1 with SVG2 as SVG2 references it.

  Cheers,

  Rich


  Rich Schwerdtfeger

  Inactive hide details for Amelia Bellamy-Royds ---05/21/2015 12:21:44
  PM---Thanks all for feedback. Léonie:Amelia Bellamy-Royds ---05/21/2015
  12:21:44 PM---Thanks all for feedback. Léonie:

  From: Amelia Bellamy-Royds <amelia.bellamy.royds@gmail.com>
  To: "public-svg-a11y@w3.org" <public-svg-a11y@w3.org>
  Cc: Fred Esch/Arlington/IBM@IBMUS, "White, Jason J" <jjwhite@ets.org>,
  Léonie Watson <lwatson@paciellogroup.com>
  Date: 05/21/2015 12:21 PM
  Subject: Re: ARIA roles for graphics



  Thanks all for feedback.

  Léonie:
  I am definitely open to alternate names. Really, we want something that
  authors will know how to use without having to double-check the spec, so
  "compleximg" or "structuredimg" make sense (even if they are somewhat
  verbose).  I'd be cautious about using more specific words, like
  "infographic", just yet.  As Fred said, we don't want to lock in anything
  that might become redundant when we get into specifics for charts.  I
  also don't want to restrict the semantic uses -- this could be used for
  fun interactive graphics, not only informational ones!

  If you don't think "figure" maps well to "complementary", then we could
  use one of the other section types.  Maybe region?

  Fred:
  I agree with your concern, but don't think it's necessary.  Even if we
  have specific roles for charts and infographics, we still need basic
  terms for other uses of SVG.

  Jason:
  I agree that testing & implementation is important.  If that means that
  we can't meet a timeline for ARIA 1.1, so be it.  Still better to get
  started sooner rather than later!

  If we have time for this on tomorrow's call, we can talk more.
  Otherwise, we can put it on the agenda for next week.

  Best,
  Amelia